Method and apparatus for generating content using response-based transfer of operation rules between separated accounts

The method of generating operational rules from reaction-based content transfer between separated accounts addresses the challenge of minimizing reputation and sanction risks in social media content creation by applying verified rules to a second account, enhancing strategy exploration and risk reduction.

KR103043685B1Active Publication Date: 2026-11-06EN HANCE
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
KR1020260130203
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2026-07-15
Publication Date
2026-11-06
Estimated Expiration
2046-07-15

AI Technical Summary

Technical Problem

Conventional content creation methods on social media platforms fail to effectively explore content strategies while minimizing reputation and sanction risks, as they directly post unverified content, leading to potential account sanctions and difficulty in translating experimental findings into operational knowledge.

Method used

A method and apparatus for creating content using reaction-based operational rule transfer between separated accounts, where operational rules are generated based on the reaction results of content posted in a first account and applied to a second account without copying the original content, incorporating both quantitative and policy risk signals to generate transition and avoidance rules.

Benefits of technology

This approach allows for exploring content strategies while reducing the risk of sanctions on the second account by converting successful content characteristics into transition rules and risky content characteristics into avoidance rules, ensuring only verified rules are applied, thus protecting the account's reputation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112026086162005-PAT00002_ABST
    Figure 112026086162005-PAT00002_ABST
Patent Text Reader

Abstract

The present invention relates to a method and apparatus for generating content using reaction-based operational rule transfer between separated accounts. A content generating apparatus according to one embodiment of the present invention includes a reaction result acquisition unit for acquiring a reaction result for a first content posted through a first account, an operational rule generation unit for generating an operational rule based on the reaction result, and a content generating unit for generating a second content to be posted through a second account separated from the first account based on the operational rule without copying the first content. The operational rule is a rule derived from the first content that defines the method of generating the second content. Accordingly, a content strategy can be explored while lowering the reputation risk of the second account, and unverified expressions can be prevented from spreading to the second account.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to a method and apparatus for creating content using reaction-based operational rule transfer between separated accounts, and more specifically, to a method and apparatus for creating operational rules based on the reaction results of content posted through a first account and applying said operational rules to the creation of content of a second account. Background Technology

[0003] When operating accounts on platforms such as social media, it is required to continuously monitor content performance, user reactions, policy risk signals, and changes in account status, and to reflect this in operational strategies.

[0004] Traditionally, a method was used to explore content strategies by posting multiple content variations within the same account and comparing their response rates. However, since this method involves directly posting unverified content on the account under management, inappropriate expressions or failed content can directly affect the account's reputation.

[0005] In addition, if policy risks are directly executed in an operational account through representations or behaviors that have not been sufficiently verified, the risk of sanctions, such as reporting, deletion, exposure restrictions, or changes in account status, may directly occur to that account.

[0006] On the other hand, conventional content experiments tend to accumulate results solely in quantitative metrics such as views or response rates, making it difficult to translate experimental findings into generalized operational knowledge. Consequently, it is difficult to sufficiently explore and accumulate information on which combinations of topics, styles, lengths, or behaviors are effective, and furthermore, which expressions entail policy risks.

[0007] Furthermore, conventional methods are primarily used to reinforce patterns of content with favorable response results, but they fail to convert characteristics of cases with poor response results or identified policy risks into avoidance rules. As a result, it is difficult to accumulate data-based rules to protect operational accounts.

[0008] Therefore, technology is required to verify content strategies and policy risk characteristics while lowering reputation and sanction risks of the accounts under operation, and to reflect the verified results in the creation of content for the said accounts by transferring them in the form of operational rules rather than copying the original content. Prior art literature

[0010] Korean Registered Patent No. 10-1657991 (September 9, 2016), Online Communication Account Operating System The problem to be solved

[0011] The objective of the present invention is to provide a method and apparatus for content creation using reaction-based operational rule transfer between separated accounts, which can explore content strategies while lowering the reputation risk of the second account by applying operational rules generated based on the reaction results of content posted in the first account to the content creation of the second account.

[0012] Another objective of the present invention is to provide a method and apparatus for creating content using reaction-based operational rule transfer between separated accounts, which can reduce the risk of sanctions on a second account by evaluating the reaction results of content including not only performance but also policy risk signals, and generating operational rules based on the evaluation results.

[0013] Another objective of the present invention is to provide a method and apparatus for creating content using reaction-based operational rule transfer between separated accounts, wherein an operational rule is generated and transferred based on the reaction result of the content without copying the original text of the content posted in the first account, and after verifying whether the operational rule satisfies a predetermined standard, it is applied to the creation of content in the second account so that only the verified operational rule is transferred.

[0014] However, the technical problems that the present invention aims to solve are not limited to those described above, and other unmentioned problems can be clearly understood by those skilled in the art from the description of the invention below. means of solving the problem

[0016] According to one embodiment of the present invention for achieving the above objective, a content generation device is disclosed comprising: a reaction result acquisition unit for acquiring a reaction result for a first content posted through a first account; an operation rule generation unit for generating an operation rule based on the reaction result; and a content generation unit for generating a second content to be posted through a second account separated from the first account based on the operation rule without copying the first content, wherein the operation rule is a rule derived from the first content and stipulating a method of generating the second content.

[0017] In one embodiment, the response result includes a quantitative response indicator and a policy risk signal for the first content, and the policy risk signal may include at least one of a report, deletion, exposure restriction, warning, and account status change.

[0018] In one embodiment, the operation rule generation unit classifies the first content into either a transition target or an avoidance target based on the response result, generates a transition rule used for generating the second content when classified as a transition target, and generates an avoidance rule that restricts the generation of the second content when classified as an avoidance target.

[0019] In one embodiment, the avoidance rule is derived from at least one of the negative reaction element and avoidance condition of the first content and may include at least one of a prohibited expression, a prohibited topic, an expression intensity upper limit, and a proximity of a prohibited expression.

[0020] In one embodiment, the transition rule is derived from at least one of the topic pattern, stylistic features, and response-inducing elements of the first content, and can define at least one of the topic, style, and response method of the second content.

[0021] In one embodiment, the content generation device further includes a rule verification unit that verifies whether the operating rule satisfies a predetermined standard, and the content generation unit can generate the second content based on the operating rule that passed the verification.

[0022] In one embodiment, the operation rule generation unit may evaluate the response result by using a quantitative response indicator for the first content and a natural language response for the first content together as input, and generate the operation rule based on the evaluation.

[0023] In one embodiment, the operation rule generation unit calculates a promotion score by using the degree of reaction of the first content as an increasing factor and at least one of the degree of policy risk of the first content and the sample uncertainty of the reaction result as a decreasing factor, and determines whether to classify the first content as a transition target based on the promotion score.

[0024] In one embodiment, the operation rule may be stored as a transition record including at least one of an applicable subject, an applicable segment, a content format, and an expiration condition.

[0025] In one embodiment, the reaction result acquisition unit further acquires a reaction result for the second content posted through the second account, and the operation rule generation unit can update the operation rule based further on the reaction result for the second content.

[0026] According to one embodiment of the present invention, to achieve the above objective, a method for creating content is disclosed, comprising: a step of obtaining a reaction result for a first content posted through a first account; a step of generating an operation rule based on the reaction result; and a step of generating a second content to be posted through a second account separated from the first account based on the operation rule without copying the first content, wherein the operation rule is a rule derived from the first content and stipulating a method for generating the second content.

[0027] In one embodiment, the response result includes a quantitative response indicator and a policy risk signal for the first content, and the policy risk signal may include at least one of a report, deletion, exposure restriction, warning, and account status change.

[0028] In one embodiment, the step of generating the operation rule may include: classifying the first content into either a transition target or an avoidance target based on the reaction result; generating a transition rule used for generating the second content when classified as a transition target; and generating an avoidance rule that restricts the generation of the second content when classified as an avoidance target.

[0029] In one embodiment, the avoidance rule is derived from at least one of the negative reaction element and avoidance condition of the first content and may include at least one of a prohibited expression, a prohibited topic, an expression intensity upper limit, and a proximity of a prohibited expression.

[0030] In one embodiment, the transition rule is derived from at least one of the topic pattern, stylistic features, and response-inducing elements of the first content, and can define at least one of the topic, style, and response method of the second content.

[0031] In one embodiment, the content generation method further includes a step of verifying whether the operating rule satisfies a predetermined standard, and the step of generating the second content may include a step of generating the second content based on the operating rule that passed the verification.

[0032] In one embodiment, the step of generating the operation rule may include evaluating the response result using a quantitative response indicator for the first content and a natural language response for the first content together as input, and generating the operation rule based on the evaluation.

[0033] In one embodiment, the step of generating the operation rule may include: a step of calculating a promotion score by using the degree of reaction of the first content as an increasing factor and at least one of the degree of policy risk of the first content and the sample uncertainty of the reaction result as a decreasing factor; and a step of determining whether to classify the first content as a transition target based on the promotion score.

[0034] In one embodiment, the operation rule may be stored as a transition record including at least one of an applicable subject, an applicable segment, a content format, and an expiration condition.

[0035] In one embodiment, the content creation method may further include the step of obtaining a reaction result for the second content posted through the second account; and the step of updating the operation rule based further on the reaction result for the second content. Effects of the invention

[0037] According to one embodiment of the present invention, by generating an operating rule based on the reaction results to a first content posted through a first account and generating a second content of a second account separated from the first account based on the operating rule, it is possible to explore a content strategy while lowering the reputation risk of the second account by preventing unverified content from being directly posted to the second account. Furthermore, since the second content is generated based on an operating rule derived from the first content without copying the first content, only generalized operating rules can be transferred while preventing unverified expressions from being propagated to the second account as is.

[0038] According to one embodiment of the present invention, the response result includes not only quantitative response indicators but also policy risk signals such as reporting, deletion, exposure restriction, warning, or account status change. By classifying the first content into transition targets and avoidance targets and generating transition rules and avoidance rules, the risk of sanctions on the second account can be reduced by learning not only the characteristics of high-performance content but also the characteristics of content involving policy risks. Accordingly, characteristics when the response result is good are converted into transition rules, and characteristics when the response result is poor or risk is confirmed are converted into avoidance rules, thereby allowing both success patterns and risk patterns to be reflected in the operation of the second account.

[0039] According to one embodiment of the present invention, after verifying whether the operating rule satisfies a predetermined standard, the second content is generated based on the operating rule that has passed the verification, and by determining whether it is subject to transfer based on promotion scores and sample uncertainty, etc., only the operating rule with secured reliability can be reflected in the second account. Furthermore, by utilizing the reaction result to the second content again for updating the operating rule, the acquisition of the reaction result, the generation of the operating rule, and the generation of the content are repeated, thereby providing a content generation device and method in which the operating rule continuously self-evolves.

[0040] However, the effects obtainable through the present invention are not limited to those described above, and other unmentioned technical effects will be clearly understood by those skilled in the art from the description of the invention below. Brief explanation of the drawing

[0042] FIG. 1 is a diagram showing a system environment to which a content creation device related to one embodiment of the present invention is applied. FIG. 2 is a block diagram showing the configuration of a content creation device related to one embodiment of the present invention. FIG. 3 is a flowchart illustrating a content creation method related to an embodiment of the present invention. FIG. 4 is a block diagram showing the detailed configuration of an operation rule generation unit related to an embodiment of the present invention. FIG. 5 is a flowchart illustrating the process of creating and updating operating rules related to an embodiment of the present invention. FIG. 6 is a diagram showing the state transition of the first content and operation rule related to one embodiment of the present invention. FIG. 7 is a block diagram showing the hardware configuration of a content creation device related to one embodiment of the present invention. FIG. 8 is a diagram showing the data structure of a transition record related to one embodiment of the present invention. Specific details for implementing the invention

[0043] An embodiment of the present invention will be described in detail below with reference to the attached drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings. Instead, based on the principle that the inventor can appropriately define the concept of a term to best describe their invention, they should be interpreted in a meaning and concept consistent with the technical spirit of the present invention. Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely partial embodiments of the present invention and do not represent all of the technical spirit of the present invention. It should be understood that various equivalents and modifications capable of replacing them may exist at the time of filing this application.

[0044] Additionally, as used herein, “comprise, include” and / or “comprising, including” specify the presence of the mentioned features, numbers, steps, actions, parts, elements, and / or groups thereof, and do not exclude the presence or addition of one or more other features, numbers, actions, parts, elements, and / or groups.

[0045] Additionally, to aid in understanding the invention, the attached drawings are not drawn to actual scale, and the dimensions of some components may be exaggerated. Furthermore, the same reference numerals may be assigned to identical components in different embodiments.

[0046] The statement that two subjects of comparison are 'identical' means that they are 'substantially identical.' Therefore, substantial identity may include deviations considered low in the industry, for example, deviations within 5%. Additionally, the statement that a parameter is uniform in a given area may mean that it is uniform from an average perspective.

[0047] Although terms such as "first," "second," etc., are used to describe various components, it goes without saying that these components are not limited by these terms. These terms are used merely to distinguish one component from another, and unless specifically stated otherwise, the first component may also be the second component.

[0048] Throughout the specification, unless specifically stated otherwise, each component may be singular or plural.

[0049] The fact that any configuration is placed on the "upper (or lower)" of a component or on the "upper (or lower)" of a component may mean not only that any configuration is placed in contact with the upper (or lower) surface of said component, but also that another configuration may be interposed between said component and any configuration placed on (or below) said component.

[0050] Furthermore, where one component is described as being "on," "connected to," or "coupled to" another component, it should be understood that while the components may be directly connected or coupled to each other, another component may be "interposed" between each component, or that each component may be "connected," "coupled," or "coupled" through another component.

[0051] As used herein, the term “and / or” includes any and all combinations of one or more of the associated listed items. Additionally, when describing embodiments of the invention, the use of “may” relates to “one or more embodiments of the invention.” Expressions such as “one or more” preceding a list of elements modify the entire list of elements and do not modify individual elements of the list.

[0052] Throughout the specification, "A and / or B" means A, B, or A and B unless specifically stated otherwise, and "C to D" means C or more and D or less, unless specifically stated otherwise.

[0053] When syntax such as "at least one of A, B, and C", "at least one of A, B, or C", "at least one selected from the group of A, B, and C", or "at least one selected from A, B, and C" is used to specify a list of elements A, B, and C, the syntax can refer to any suitable combination.

[0054] The term "use" may be considered synonymous with the term "utilize." As used herein, terms such as "substantially," "about," and similar terms are used as terms of approximation rather than degree, and are intended to account for the inherent variation of a measured or calculated value that would be recognized by a person skilled in the art.

[0055] In this specification, terms such as first, second, third, etc. may be used to describe various elements, components, regions, layers, and / or sections, but these elements, components, regions, layers, and / or sections should not be limited by these terms. These terms are used to distinguish one element, component, region, layer, or section from another element, component, region, layer, or section. Accordingly, the first element, component, region, layer, or section discussed below may be named the second element, component, region, layer, or section without departing from the teachings of the exemplary embodiments.

[0056] Spatial relative terms such as "beneath," "below," "lower," "above," and "upper" may be used in the specification for ease of description to explain the relationship between one element or feature and other element(s) or feature(s) as illustrated in the drawings. Spatially relative positions are to be understood as encompassing different orientations of the device in use or operation, in addition to the orientations depicted in the drawings. For example, if the device in the drawings is inverted, the other element is understood as "below" or "below," and the described element is understood as "above" or "upper" of the other element. Thus, the term "below" may encompass both the up and down directions.

[0057] The terms used in this specification are for describing embodiments of the invention and are not intended to limit the invention.

[0058] In this specification, the term “platform” may mean a service through which content can be posted via an account. A platform may be a social media service, a community service, a messaging service, or a content sharing service, and is not limited to a specific service.

[0059] In this specification, the term "content" may refer to a unit of information that can be posted through a platform. In one embodiment, the content may include at least one of a post, a comment on the post, a reaction to the post, sharing, saving, and an equivalent interaction. The reaction may include a like, a recommendation, a bookmark, or a corresponding mark. Accordingly, the content may include text, images, video, audio, or a combination thereof, and is not limited to a specific format.

[0060] In this specification, "first account" and "second account" may refer to accounts that are operated separately from each other on the platform. In one embodiment, the first account may be an account that does not represent the second account, or an account whose identification information is distinct from that of the second account. That is, it may be an account in which the results of posting in the first account do not directly affect the reputation or account status of the second account. The first account and the second account may each be one or multiple. In one embodiment, the first account may be an experimental account for limited verification of content, response methods, behavior combinations, or policy risk characteristics, and the second account may be a main account representing the brand.

[0061] In this specification, "first content" may refer to content posted through a first account, and "second content" may refer to content posted through a second account. The first content and the second content may each include one or a plurality of contents. That is, "first" and "second" merely indicate a distinction based on the account in which the content is posted, and do not limit the number, creation order, or posting order of the contents.

[0062] In this specification, the term "response result" may refer to information observed regarding posted content. In one embodiment, the response result may include at least one of a quantitative response indicator and a policy risk signal. The quantitative response indicator may include the number of views, impressions, clicks, saves, shares, or reactions, and the policy risk signal may include at least one of a report, deletion, exposure restriction, warning, and account status change. In this case, the exposure restriction may include hiding the content, reducing reach, and excluding it from recommendations, and the account status change may include suspending the account, restricting it, and downgrading its rating. The response result may further include natural language responses such as comments.

[0063] In this specification, the term "operational rule" may mean a rule derived from the first content that defines the method of creating the second content. The operational rule may not include the original text of the first content, thereby preventing unverified expressions from being propagated to the second account as is. In one embodiment, the operational rule may include at least one of a transition rule used for creating the second content and an avoidance rule that restricts the creation of the second content.

[0064] FIG. 1 is a diagram showing a system environment to which a content creation device related to one embodiment of the present invention is applied.

[0065] Referring to FIG. 1, a content creation device (100) can be linked with a platform (200). A first account (210) and a second account (220) separated from the first account (210) can be opened on the platform (200).

[0066] Although FIG. 1 illustrates that the first account (210) and the second account (220) are opened on a single platform (200), the first account (210) and the second account (220) may each be opened on different platforms. In this case, the platform (200) may be understood as a collective term for multiple platforms. Additionally, the first account (210) may be multiple, and in this case, the multiple first accounts may each be responsible for different segments, languages, regions, or tones.

[0067] The content creation device (100) can obtain a reaction result for the first content posted through the first account (210). The content creation device (100) can create an operation rule based on the obtained reaction result, and can create a second content to be posted through the second account (220) based on the created operation rule. At this time, the second content can be created without copying the first content.

[0068] Since the first account (210) is operated separately from the second account (220), even if the first content posted through the first account (210) receives a low response or entails policy risks, the reputation risk and sanction risk of the second account (220) may not be directly increased. Accordingly, the content strategy and policy risk characteristics can be verified without increasing the risk of the second account (220).

[0069] In addition, since the second content is created based on the operating rules derived from the first content and is not created by copying the original text of the first content, it is possible to prevent unverified expressions from being propagated to the second account (220) as is.

[0070] In one embodiment, the first account (210) may be operated according to certain restrictions. The restrictions may relate to at least one of the activity volume, posting period, posting frequency, access scope to sensitive topics, and promotional intensity of the first account (210). Verification in the first account (210) may be performed within a controlled range according to the restrictions.

[0071] The content creation device (100) may post the created second content through the second account (220). However, the content creation device (100) is not required to perform the posting of the second content; the content creation device (100) may create the second content, and the posting may be performed by a separate device or operator.

[0072] The content creation device (100) can further obtain reaction results for the second content posted through the second account (220). The obtained reaction results for the second content can be used to update the operation rules. Accordingly, the acquisition of reaction results, the creation of operation rules, and the creation of content can be performed repeatedly, and the operation rules can be continuously updated.

[0073] In one embodiment, the updated operating rule may be used not only for the creation of the second content but also for the creation of the first content to be subsequently posted through the first account (210). In this case, the updated operating rule may be used to limit the search range for the subject, style, or intensity of expression of the subsequent first content, and accordingly, verification in the first account (210) may be performed reflecting the results of the previous verification.

[0074] The content creation device (100) can obtain response results using at least one of an application programming interface (API) provided by the platform (200), observation through a browser, an analysis tool, manual input, and a third-party data source.

[0075] The content creation device (100) may be implemented as a server, a user terminal, a cloud system, or a combination thereof, and is not limited to a specific physical form.

[0076] FIG. 2 is a block diagram showing the configuration of a content creation device related to one embodiment of the present invention.

[0077] Referring to FIG. 2, the content generation device (100) may include a response result acquisition unit (110), an operation rule generation unit (120), and a content generation unit (140). The content generation device (100) may further include a rule verification unit (130).

[0078] The components illustrated in FIG. 2 are not essential for implementing the content creation device (100), and the content creation device (100) may have more or fewer components than those listed above.

[0079] The reaction result acquisition unit (110) can acquire a reaction result for the first content posted through the first account (210).

[0080] The reaction result obtained by the reaction result acquisition unit (110) may include a quantitative reaction indicator for the first content. The quantitative reaction indicator may include the number of views, impressions, clicks, saves, shares, or reaction indicators, and may further include reactions by time period, reactions by follower segment, or relative performance compared to similar content.

[0081] Additionally, the reaction result obtained by the reaction result acquisition unit (110) may include a policy risk signal for the first content. The policy risk signal may include at least one of a report, deletion, exposure restriction, warning, and account status change. That is, the reaction result may include not only information indicating how excellent the performance of the first content was, but also information indicating what sanction risk the first content entailed.

[0082] The reaction result obtained by the reaction result acquisition unit (110) may further include natural language reactions to the first content. The natural language reactions may include comments, and the comments may include expressions indicating affirmation, negation, question, rebuttal, or confusion.

[0083] The operation rule generation unit (120) can generate an operation rule based on the reaction result obtained by the reaction result acquisition unit (110). The operation rule may be a rule derived from the first content that defines the method of generating the second content.

[0084] The operation rule generation unit (120) can evaluate the response results by using a quantitative response indicator for the first content and a natural language response for the first content together as input, and can generate an operation rule based on the evaluation.

[0085] For example, even if the first content obtains a high quantitative response indicator, if a negative natural language response is also observed, the operation rule generation unit (120) can generate an operation rule so that the first content is not used to generate the second content.

[0086] The operation rule generation unit (120) can classify the first content into either a transition target or an avoidance target based on the response result. If the first content is classified as a transition target, the operation rule generation unit (120) can generate a transition rule used for generating the second content, and if the first content is classified as an avoidance target, it can generate an avoidance rule that restricts the generation of the second content.

[0087] That is, the operation rule generation unit (120) can use not only the characteristics of the first content with a good response result, but also the characteristics of the first content with a poor response result or confirmed policy risk as an avoidance rule. The detailed configuration of the operation rule generation unit (120) will be described later with reference to FIG. 4.

[0088] In one embodiment, the operation rule generation unit (120) can calculate a promotion score by using the degree of reaction of the first content as an increase factor and at least one of the degree of policy risk of the first content and the sample uncertainty of the reaction result as a decrease factor. The operation rule generation unit (120) can determine whether to classify the first content as a transition target based on the calculated promotion score.

[0089] In one embodiment, the operation rule generation unit (120) may classify the first content as a pending or human review subject rather than a transition subject if at least one of the conditions of the minimum sample size, observation period, segment consistency, and absence of a risk signal is not satisfied even when the promotion score is greater than or equal to a threshold value. Accordingly, it is prevented that an operation rule that has not been sufficiently verified is used to generate the second content.

[0090] The rule verification unit (130) can verify whether the operation rule generated by the operation rule generation unit (120) satisfies a predetermined standard. The predetermined standard may be at least one of brand tone suitability, legal risk, regulatory risk, sensitive subject risk, community policy risk, conflict with existing prohibited subjects, promotional intensity, sample size, and data reliability.

[0091] The rule verification unit (130) may classify operation rules that have not passed the verification as pending, discarded, or subject to human review. Accordingly, the operation rules generated by the operation rule generation unit (120) are not immediately applied to the creation of the second content, but only the operation rules that have passed the verification can be applied to the creation of the second content.

[0092] In one embodiment, the rule verification unit (130) may not perform the verification on the avoidance rule. This is because the avoidance rule is a rule that restricts the creation of the second content, so even if it is applied without verification, it may not increase the risk of the second account (220).

[0093] The content creation unit (140) can create second content to be posted through a second account (220) separated from the first account (210) based on operating rules without copying the first content. If the content creation device (100) includes a rule verification unit (130), the content creation unit (140) can create second content based on operating rules that have passed the verification.

[0094] The content generation unit (140) can determine at least one of the subject, style, and response method of the second content based on transition rules, and can determine an expression or subject that must not be included in the second content based on avoidance rules.

[0095] The content generation unit (140) can generate a second content using a generative artificial intelligence model. In this case, the operation rule can be used to update at least one of a prompt input to the generative artificial intelligence model, a template applied to the prompt, a forbidden word filter, a response mode weight, a behavior portfolio probability, a time-based schedule weight, and a criterion for approval.

[0096] The second content generated by the content generation unit (140) can be posted through the second account (220). The reaction result acquisition unit (110) can further acquire reaction results for the second content posted through the second account (220), and the operation rule generation unit (120) can further update the operation rules based on the reaction results for the second content.

[0097] That is, as illustrated in FIG. 2, the reaction result for the second content generated by the content generation unit (140) is input back into the reaction result acquisition unit (110), thereby forming a closed loop in which the acquisition of the reaction result, the generation of the operation rule, and the generation of the second content are repeatedly performed.

[0098] In one embodiment, the update of the operating rules may be performed automatically, semi-automatically, or after human approval.

[0099] The response result acquisition unit (110), operation rule generation unit (120), rule verification unit (130), and content generation unit (140) may be physically separated separate components, or they may be functional components implemented by one or more processors. This will be described later with reference to FIG. 7.

[0100] FIG. 3 is a flowchart illustrating a content creation method related to an embodiment of the present invention.

[0101] The content creation method illustrated in FIG. 3 can be performed by the content creation device (100) described with reference to FIG. 2. Therefore, even if details are omitted below, the details described with respect to FIG. 1 and FIG. 2 can also be applied to the content creation method illustrated in FIG. 3.

[0102] The content creation device (100) can obtain a reaction result for the first content posted through the first account (210) (S310).

[0103] The above response result may include quantitative response indicators and policy risk signals for the first content, and the policy risk signal may include at least one of reporting, deletion, exposure restriction, warning, and account status change. The above response result may further include natural language responses such as comments.

[0104] The content creation device (100) can evaluate the obtained reaction results (S320).

[0105] The content generation device (100) can evaluate the response result by using a quantitative response indicator for the first content and a natural language response for the first content together as inputs.

[0106] In one embodiment, the content generating device (100) can calculate at least one of the reaction level of the first content, the policy risk level of the first content, and the sample uncertainty of the reaction result as the evaluation. The content generating device (100) can calculate a promotion score by using the reaction level as an increase factor and at least one of the policy risk level and the sample uncertainty as a decrease factor.

[0107] The content creation device (100) can determine whether the first content is a transfer target based on the above evaluation (S330).

[0108] The content creation device (100) can determine whether to classify the first content as a transition target based on the calculated promotion score.

[0109] In one embodiment, the content generating device (100) may classify the first content as a pending or human review subject rather than classifying it as a transition subject if at least one of the conditions of the minimum sample size, observation period, segment consistency, and absence of a risk signal is not satisfied even when the promotion score is greater than or equal to a reference value.

[0110] If the first content is classified as a transition target, the content creation device (100) can create a transition rule used to create the second content (S340).

[0111] The above transition rule may be derived from at least one of the topic pattern, stylistic features, and response-inducing elements of the first content, and may define at least one of the topic, style, and response method of the second content.

[0112] If the first content is not classified as a transition target, that is, if the first content is classified as an avoidance target, the content creation device (100) can create an avoidance rule that restricts the creation of the second content (S350).

[0113] The above avoidance rule may be derived from at least one of the negative reaction element and avoidance condition of the first content, and may include at least one of a prohibited expression, a prohibited topic, an upper limit on expression intensity, and a proximity to a prohibited expression.

[0114] In this way, the content generation device (100) can convert and utilize not only the characteristics of the first content with good response results, but also the characteristics of the first content with poor response results or confirmed policy risks into avoidance rules. Accordingly, rules for protecting the second account (220) can be accumulated based on data.

[0115] The content creation device (100) can verify whether the generated operation rule satisfies a predetermined standard and determine whether the operation rule has passed the verification (S360).

[0116] The above-mentioned criteria may relate to at least one of brand tone suitability, legal risk, sensitive subject risk, community policy risk, sample size, and data reliability.

[0117] If the above operating rule does not pass the above verification, the content creation device (100) may classify the above operating rule as pending or subject to human review (S370).

[0118] Operational rules classified as pending or subject to human review may not apply to the creation of the second content, may be discarded, or may apply to the creation of the second content after human approval.

[0119] If the above operating rule passes the above verification, the content creation device (100) can create a second content to be posted through a second account (220) separated from the first account (210) based on the above operating rule without copying the first content (S380).

[0120] The content generating device (100) can determine at least one of the subject, style, and response method of the second content based on transition rules, and can determine an expression or subject that must not be included in the second content based on avoidance rules.

[0121] Figure 3 illustrates that verification of step S360 is performed even when an avoidance rule is generated, but in one embodiment, the avoidance rule may be applied to the creation of the second content without undergoing verification of step S360.

[0122] The content creation device (100) can further obtain reaction results for the second content posted through the second account (220) (S390).

[0123] The reaction result to the second content obtained in step S390 may be subject to evaluation in step S320. That is, the content generating device (100) may update the operation rule by repeatedly performing steps S320 through S380 on the reaction result to the second content. In this way, the update of the operation rule is not performed by a separate, independent step, but can be achieved by repeatedly performing the above steps targeting the reaction result to the second content.

[0124] Accordingly, a closed loop can be formed in which the acquisition of response results, evaluation of response results, creation of operation rules, and creation of second content are repeatedly performed, and the operation rules can be continuously updated.

[0125] In one embodiment, the updated operating rule is used not only for the creation of the second content, but also for the creation of the first content to be subsequently posted through the first account (210). This will be described later with reference to FIG. 5.

[0126] Each step illustrated in FIG. 3 does not necessarily have to be performed in the order illustrated, some steps may be performed simultaneously, the order may be changed, and some steps may be omitted.

[0127] FIG. 4 is a block diagram showing the detailed configuration of an operation rule generation unit related to an embodiment of the present invention.

[0128] Referring to FIG. 4, the operation rule generation unit (120) may include an evaluation unit (121), a classification unit (122), a transition rule generation unit (123), and an avoidance rule generation unit (124).

[0129] The components illustrated in FIG. 4 are not essential for implementing the operation rule generation unit (120), and the operation rule generation unit (120) may have more or fewer components than those listed above. For example, the evaluation unit (121) and the classification unit (122) may be implemented as a single integrated component, and the transition rule generation unit (123) and the avoidance rule generation unit (124) may also be implemented as a single integrated component.

[0130] The evaluation unit (121) can receive the reaction result from the reaction result acquisition unit (110) and can evaluate the received reaction result.

[0131] The evaluation unit (121) can evaluate the response results by using quantitative response indicators for the first content and natural language responses for the first content together as input. That is, the evaluation unit (121) can evaluate the response results by using not only quantitative information such as the number of views, impressions, clicks, saves, shares, or response indicators, but also qualitative information such as expressions indicating positivity, negativity, questions, rebuttals, or confusion included in the comments.

[0132] Accordingly, even if the first content obtains high quantitative response indicators, if it is evaluated as having a high negative natural language response, potential for policy violation, or brand non-conformity, the first content may not be used to generate the second content.

[0133] In one embodiment, the evaluation unit (121) can calculate at least one of the reaction level of the first content, the brand suitability level of the first content, the policy risk level of the first content, the negative reaction level of the first content, the segment consistency, and the sample uncertainty of the reaction result as a result of the evaluation.

[0134] In one embodiment, the evaluation unit (121) may calculate a promotion score by using the degree of reaction of the first content as an increasing factor and at least one of the degree of policy risk of the first content and the sample uncertainty of the reaction result as a decreasing factor. The promotion score may be a value indicating the possibility that the first content will be classified as a transition target.

[0135] In one embodiment, the evaluation unit (121) can evaluate the response result using an artificial intelligence model and output the result of the evaluation in a structured form. For example, the evaluation unit (121) can output structured data including an identifier of the first content, a value indicating the degree of response, a value indicating the degree of brand suitability, a value indicating the degree of policy risk, and a transition recommendation value.

[0136] The classification unit (122) can classify the first content into either a transfer target or an avoidance target based on the evaluation of the evaluation unit (121). The classification unit (122) can determine whether to classify the first content into a transfer target based on the promotion score calculated by the evaluation unit (121).

[0137] In one embodiment, the classification unit (122) may classify the first content as a pending or human review subject rather than classifying it as a transition subject if at least one of the minimum sample size, observation period, segment consistency, absence of a risk signal, and policy matching condition of the second account (220) is not satisfied, even if the promotion score is greater than or equal to a threshold value. Accordingly, it is prevented that response results for which reliability is not ensured due to insufficient samples or a short observation period are converted into operating rules.

[0138] In one embodiment, the classification unit (122) may classify the first content into one of a transition target, an avoidance target, and a hold target. The avoidance target may include the first content with a poor response result and the first content with a confirmed policy risk, and the hold target may include the first content with insufficient samples of response results. The first content classified as a hold target may be reclassified after additional response results are obtained. This will be described later with reference to FIG. 6.

[0139] The transition rule generation unit (123) can generate a transition rule used to generate the second content when the first content is classified as a transition target.

[0140] The transition rule generated by the transition rule generation unit (123) can be derived from at least one of the topic pattern, stylistic features, and response-inducing elements of the first content. Additionally, the transition rule can define at least one of the topic, style, and response method of the second content.

[0141] The transition rule generation unit (123) can generate a transition rule that does not include the original text of the first content as is. For example, the transition rule generation unit (123) can generate a rule that a combination of a specific topic and a specific writing style induces a high response, a rule that a question-type introduction is effective for content regarding technology comparison, or a rule that a question-type hooking is effective when connecting a specific trend and a specific topic.

[0142] In this way, the transfer rule generation unit (123) generates a generalized rule rather than the original text of the first content, so that the second account (220) does not imitate the first content posted on the first account (210) exactly as it is, and only the generalizable rule can be transferred.

[0143] The avoidance rule generation unit (124) can generate an avoidance rule that restricts the generation of the second content when the first content is classified as an avoidance target.

[0144] The avoidance rule generated by the avoidance rule generation unit (124) can be derived from at least one of the negative reaction element and avoidance condition of the first content. Additionally, the avoidance rule may include at least one of a prohibited expression, a prohibited topic, an expression intensity upper limit, and a proximity to a prohibited expression.

[0145] For example, the avoidance rule generation unit (124) may generate an avoidance rule that negative reactions increase above a certain expression intensity, an avoidance rule that expressions close to a certain prohibition condition must be avoided in the second account (220), or an avoidance rule that specific topics must not be mentioned in the second account (220).

[0146] In one embodiment, the avoidance rule generated by the avoidance rule generation unit (124) can be converted into an approval standard that restricts the automatic posting of the second content. That is, if there is a possibility that an expression corresponding to the avoidance rule is included in the second content, the second content may not be automatically posted but may be posted after receiving human approval.

[0147] The transition rule generated by the transition rule generation unit (123) and the avoidance rule generated by the avoidance rule generation unit (124) can be transmitted to the rule verification unit (130). However, in one embodiment, the avoidance rule may be transmitted to the content generation unit (140) without passing through the rule verification unit (130).

[0148] In this way, the operation rule generation unit (120) can generate a transition rule from the first content with a good response result and generate an avoidance rule from the first content with a poor response result or a confirmed policy risk, thereby reflecting both the success pattern and the risk pattern in the operation of the second account (220).

[0149] FIG. 5 is a flowchart illustrating the process of creating and updating operating rules related to one embodiment of the present invention.

[0150] FIG. 5 shows a specific embodiment in which the content creation method described with reference to FIG. 3 is applied to a verification environment using the first account (210). Therefore, even if details are omitted below, the details described with respect to FIG. 1 to FIG. 4 can also be applied to the process illustrated in FIG. 5.

[0151] In the following description, it is assumed that the first account (210) is an experimental account for limited verification of content, response methods, behavior combinations, or policy risk characteristics, and the second account (220) is the main account representing the brand. However, this is merely an example to aid understanding, and the present invention is not limited thereto.

[0152] The content creation device (100) can design a candidate for the first content to be posted through the first account (210) (S510). Hereinafter, the process of designing the candidate is referred to as experimental design.

[0153] A content generation device (100) can design a candidate for a first content by combining at least one of the content's subject, style, length, format, response type, behavior combination, and policy risk characteristics. The subject may include a technical subject, an industry issue, a customer problem, a trend, or a case, and the style may include a professional style, a casual style, a questioning style, a rebuttal style, a summary style, or an experience-linking style. The format may include a short sentence, a list style, a questioning style, an image-accompanied style, a long text, or a comment, and the behavior combination may include original posting, a reply, a reaction display, author-related behavior, or a follow-up posting. The policy risk characteristics may include at least one of the distinction between direct and indirect expressions, whether sensitive topics are included, promotional intensity, assertiveness, and proximity to prohibited expressions.

[0154] The content generation device (100) can design a candidate for the first content using at least one of random search, hypothesis-based design, multi-slot bandit, Bayesian optimization, and candidate generation by an artificial intelligence model.

[0155] The content creation device (100) can execute a candidate for the designed first content through the first account (210) (S520). Below, the execution may include publishing the first content through the first account (210).

[0156] The content creation device (100) may apply at least one restriction condition among a per-post budget, time slot, action limit, and risk limit to the first account (210). Additionally, the first content posted through the first account (210) may not contain the same sentence as the content posted through the second account (220).

[0157] The content creation device (100) can collect performance signals and risk signals for the first content (S530).

[0158] The above performance signal may include at least one of views, impressions, clicks, saves, shares, and reactions, and the above risk signal may include at least one of reports, deletions, exposure restrictions, warnings, and account status changes. The content creation device (100) may further collect at least one of reactions by time period, reactions by author or follower segment, and relative performance compared to similar content.

[0159] The performance signal and risk signal collected in step S530 can correspond to the reaction result obtained in step S310 of Fig. 3.

[0160] The content creation device (100) can determine whether the first content meets a predetermined promotion criterion based on collected performance signals and risk signals (S540).

[0161] The content generation device (100) can calculate a promotion score by using the degree of reaction, brand suitability, and segment consistency of the first content as increasing factors, and the degree of policy risk, degree of negative reaction, and sample uncertainty of the reaction result of the first content as decreasing factors.

[0162] For example, the above promotion score can be calculated by summing the value representing the degree of response, the value representing the degree of brand fit, and the value representing segment consistency, and subtracting the value representing the degree of policy risk, the value representing the degree of negative response, and the value representing sample uncertainty.

[0163] The content creation device (100) may determine that the first content satisfies the promotion criteria if the promotion score is greater than or equal to the threshold value. However, even if the promotion score is greater than or equal to the threshold value, if at least one of the minimum sample size, observation period, segment consistency, absence of risk signals, and policy matching conditions of the second account (220) is not satisfied, the content creation device (100) may determine that the first content does not satisfy the promotion criteria or classify the first content as pending or subject to human review.

[0164] The judgment of step S540 can correspond to determining whether to classify the first content as a transition target in step S330 of FIG. 3.

[0165] If it is determined that the first content does not meet the promotion criteria, the content creation device (100) may create and store an avoidance rule (S550). That is, the first content that does not meet the promotion criteria is not discarded, but can be converted into and stored as an avoidance rule that restricts the creation of the second content.

[0166] The stored avoidance rules may be used as prohibited expression filters, prohibited topic filters, expression intensity limits, or approval requirement criteria applied when creating the second content. Additionally, the stored avoidance rules may be reflected back into the experimental design of step S510, in which case the avoidance rules may be used as restriction conditions for candidates of the first content subsequently designed.

[0167] If it is determined that the first content satisfies the above promotion criteria, the content creation device (100) can abstract operation rules from the first content (S560).

[0168] The content generating device (100) may generate an operation rule that represents at least one of the topic pattern, stylistic features, response-inducing elements, and avoidance conditions of the first content, without copying the original text of the first content as is. For example, the content generating device (100) may generate an operation rule that short responsive comments are more effective than long descriptive comments in a specific segment.

[0169] The operation rule generated in step S560 can correspond to the transition rule generated in step S340 of Fig. 3.

[0170] The content creation device (100) can verify whether an abstracted operation rule satisfies a predetermined safety standard and determine whether the operation rule has passed the verification (S570). Hereinafter, the verification is referred to as a safety gate.

[0171] The above safety criteria may be at least one of brand tone suitability, legal risk, regulatory risk, sensitive subject risk, community policy risk, reputation risk of the second account (220), conflict with existing prohibited subjects, promotional intensity, data reliability, and sample size.

[0172] The judgment of step S570 can correspond to the verification of step S360 of Fig. 3.

[0173] If the above operating rule fails to pass the above safety gate, the content creation device (100) may classify the above operating rule as pending or subject to human review (S580). An operating rule classified as pending or subject to human review may not be reflected in the second account (220), may be discarded, or may be reflected after receiving human approval.

[0174] When the above operating rule passes through the above safety gate, the content creation device (100) can update the knowledge base by reflecting the above operating rule in the knowledge base (S590).

[0175] The above knowledge base may be a database in which verified operational rules are stored. Updating the above knowledge base may include updating at least one of the prompts, prompt templates, response mode weights, forbidden word filters, forbidden topic filters, behavior portfolio probabilities, time-based schedule weights, approval requirement criteria, and brand tone guidelines used for generating the second content.

[0176] The update of the above knowledge base can be performed automatically, semi-automatically, or after human approval. Additionally, the above knowledge base can be implemented in at least one form among a vector database, a rule file, a prompt template, a policy graph, and a training dataset of an artificial intelligence model.

[0177] The content creation device (100) can create a second content based on operating rules stored in an updated knowledge base and publish the created second content through a second account (220) (S600). Step S600 can correspond to step S380 of FIG. 3.

[0178] Meanwhile, the operation rules stored in the knowledge base updated in step S590 can be reflected back into the experiment design of step S510. That is, the operation rules that have passed verification and been reflected in the knowledge base can be used not only for the creation of the second content, but also for designing candidates for the first content to be subsequently posted through the first account (210).

[0179] In this case, the above operating rule may be used to limit the scope of exploration for the subject, style, or intensity of expression of the subsequent first content, and may also be used as a hypothesis for subsequent experiments. Accordingly, verification in the first account (210) may be performed by reflecting the results of previous verification.

[0180] As such, with reference to FIG. 5, the operation rule stored in the knowledge base updated in step S590 and the avoidance rule stored in step S550 can each be reflected in the experiment design of step S510. Accordingly, the content generation device (100) can design a subsequent experiment using both the operation rule derived from the first content with good response results and the avoidance rule derived from the first content with poor response results or confirmed policy risks, and a closed loop can be formed in which experiment design, publication, collection, evaluation, abstraction, verification, and reflection are repeatedly performed.

[0181] In one embodiment, the effect according to the present invention may be measured by at least one indicator among the number of first content not directly posted to the second account (220), the promotion rate of the first content, the change in performance after the application of the operation rule reflected in the knowledge base, the number of avoidance rules generated from the first content classified as avoidance targets, the number of operation rules held or blocked at the safety gate, the reduction rate of policy violations of the second account (220), the reduction rate of negative reactions of the second account (220), and the rate at which the operation rule was reused in subsequent experimental designs.

[0182] FIG. 6 is a diagram showing the state transition of the first content and operation rule related to one embodiment of the present invention.

[0183] S710 to S790 illustrated in FIG. 6 represent states that the first content or the operation rule derived from the first content may have. FIG. 6 corresponds to a specific embodiment showing how the state of the first content and the operation rule transitions during the process described with reference to FIG. 3 and FIG. 5. Therefore, even if details are omitted below, the descriptions of FIG. 1 to FIG. 5 may also apply to the state transition illustrated in FIG. 6.

[0184] The first content may have a design state (S710).

[0185] The design state may be a state where a candidate for the first content has been designed but has not yet been posted through the first account (210). The design state may correspond to step S510 of FIG. 5.

[0186] When the first content is posted through the first account (210), the first content can be transitioned from the design state to the post state (S720).

[0187] The posting status may be a state in which the first content is posted through the first account (210) and a reaction result can occur. The posting status may correspond to step S520 of FIG. 5.

[0188] When the reaction results for the first content are collected, the first content can be transitioned from the published state to the observed state (S730).

[0189] The observation state may be a state in which at least one of a quantitative response indicator, a natural language response, and a policy risk signal for the first content is collected. The observation state may correspond to step S530 of FIG. 5.

[0190] The first content in the observation state may be transferred to one of the states of transfer target (S740), avoidance target (S770), and hold (S790) depending on the evaluation of the collected reaction results.

[0191] If the first content satisfies a predetermined transition condition, the first content can be transitioned from an observation state to a transition target state (S740).

[0192] The above transition condition may include the first content having a high level of responsiveness and a low level of policy risk. In one embodiment, the above transition condition may include the first content's promotion score being greater than or equal to a threshold value.

[0193] If the first content satisfies a predetermined avoidance condition, the first content may be transitioned from an observation state to an avoidance target state (S770).

[0194] The above avoidance conditions may include at least one of the fact that the level of responsiveness of the first content is low and that a policy violation signal or a negative response signal is observed with respect to the first content.

[0195] That is, in one embodiment, the avoidance target state may include a failure state in which the level of responsiveness of the first content is low and a risk state in which a policy risk is identified for the first content. The first content in the failure state may be converted to an avoidance rule indicating an element that failed to induce a response, and the first content in the risk state may be converted to an avoidance rule indicating a prohibited expression, a prohibited topic, or a proximity criterion to a prohibited expression.

[0196] If there is insufficient data of the collected reaction results for the first content, the first content may be transitioned from an observation state to a hold state (S790).

[0197] Cases where the above data is insufficient may include at least one of cases where the sample size of the response results falls short of the minimum sample size, where the observation period is insufficient, and where segment consistency is not ensured.

[0198] First content in a pending state can be transitioned back to a published state. That is, first content in a pending state is not classified as a subject for transition or avoidance, and can be re-evaluated after additional response results are collected. Accordingly, the generation of operational rules based on response results whose reliability has not been secured can be prevented.

[0199] In one embodiment, the first content in a pending state may be classified as a subject for human review, and may be transitioned to a transition target state or an avoidance target state based on human judgment.

[0200] The operation rule derived from the first content in the transition target state can be transitioned to the rule verification state (S750).

[0201] The rule verification state may be a state in which the above-mentioned operating rule is verified to satisfy at least one of the criteria regarding brand tone suitability, legal risk, sensitive subject risk, community policy risk, sample size, and data reliability. The rule verification state may correspond to step S570 of FIG. 5.

[0202] If the above operating rule passes the above verification, the above operating rule may be transferred to a second account reflection state (S760).

[0203] The second account reflection state may be a state in which the above-mentioned operating rule is reflected in the knowledge base and can be used for the creation of the second content. The second account reflection state may correspond to step S590 of FIG. 5.

[0204] Although not illustrated in FIG. 6, if the above operating rule fails to pass the verification, the above operating rule may not be transferred to a second account reflection state and may be classified as a pending state or subject to human review.

[0205] The avoidance rule derived from the first content in the avoidance target state can be transitioned to the avoidance rule reflection state (S780).

[0206] The state in which the avoidance rule is reflected may be a state in which the above avoidance rule is reflected as a rule that restricts the creation of the second content. The above avoidance rule may be reflected as a prohibited expression filter, a prohibited topic filter, an expression intensity limit, or an approval requirement standard applied when creating the second content.

[0207] Referring to FIG. 6, an avoidance rule derived from the first content in the avoidance target state can be transitioned to the avoidance rule reflection state without going through the rule verification state. This is because the avoidance rule is a rule that restricts the creation of the second content, so even if it is reflected without verification, it may not increase the risk of the second account (220). However, in one embodiment, the avoidance rule may also be transitioned to the avoidance rule reflection state after going through the rule verification state.

[0208] According to the state transition illustrated in FIG. 6, the first content is transitioned to one of the states of transition target, avoidance target, and hold, and an operation rule used for creating the second content can be derived from the first content in the transition target state, and an avoidance rule that restricts the creation of the second content can be derived from the first content in the avoidance target state.

[0209] That is, the first content is not used only when the response result is good, but can also be converted into an avoidance rule to protect the second account (220) and used even when the response result is poor or policy risk is identified. Accordingly, both success patterns and risk patterns can be reflected in the operation of the second account (220).

[0210] FIG. 7 is a block diagram showing the hardware configuration of a content creation device related to one embodiment of the present invention.

[0211] Referring to FIG. 7, the content creation device (100) may include a processor (810), memory (820), communication unit (830), and storage unit (850), and the processor (810), memory (820), communication unit (830), and storage unit (850) may be connected to each other through a bus (840).

[0212] The components illustrated in FIG. 7 are not essential for implementing the content creation device (100), and the content creation device (100) may have more or fewer components than those listed above.

[0213] The processor (810) can control the overall operation of the content creation device (100). The processor (810) may be one or more and may be implemented as a central processing unit, a graphics processing unit, a neural network processing unit, a microprocessor, or a combination thereof.

[0214] Instructions executable by the processor (810) may be stored in the memory (820). By the processor (810) executing the instructions stored in the memory (820), the operation of the response result acquisition unit (110), operation rule generation unit (120), rule verification unit (130), and content generation unit (140) described with reference to FIG. 2 may be performed. Additionally, by the processor (810) executing the instructions, the operation of the evaluation unit (121), classification unit (122), transition rule generation unit (123), and avoidance rule generation unit (124) described with reference to FIG. 4 may be performed.

[0215] That is, the response result acquisition unit (110), operation rule generation unit (120), rule verification unit (130), and content generation unit (140) may be physically distinct components, or they may be functional components implemented by one or more processors (810). Therefore, each of the above components does not necessarily have to be implemented as separate hardware.

[0216] The memory (820) can be implemented as RAM, ROM, flash memory, or a combination thereof.

[0217] The communication unit (830) can communicate with the platform (200). The communication unit (830) can receive the result of a reaction to the first content posted through the first account (210) and the result of a reaction to the second content posted through the second account (220), and can transmit the second content generated by the content creation unit (140) to the platform (200).

[0218] The communication unit (830) can receive response results using at least one of an application programming interface (API) provided by the platform (200), observation through a browser, an analysis tool, manual input, and a third-party data source.

[0219] The storage unit (850) can store the operation rules generated by the operation rule generation unit (120). The operation rules stored in the storage unit (850) can be used by the content generation unit (140) to generate second content.

[0220] In one embodiment, the storage unit (850) may store the operation rules in the form of a transition record. The transition record will be described later with reference to FIG. 8.

[0221] In one embodiment, the storage unit (850) may correspond to the knowledge base described with reference to FIG. 5. The storage unit (850) may be implemented in at least one form among a vector database, a rule file, a prompt template, a policy graph, and a training dataset of an artificial intelligence model.

[0222] The storage unit (850) may be implemented as a separate configuration from the memory (820), or as a configuration included in the memory (820). Additionally, the storage unit (850) may be implemented as a separate database located outside the content creation device (100), in which case the content creation device (100) can access the database through the communication unit (830).

[0223] The content creation device (100) may be implemented as a server, a user terminal, a cloud system, or a combination thereof. Additionally, each part described with reference to FIGS. 2 and FIGS. 4 may be implemented distributed across multiple devices.

[0224] Meanwhile, a content generation method according to one embodiment of the present invention may be implemented in the form of a computer-readable recording medium storing instructions that cause the method to be executed when executed by one or more processors (810). The computer-readable recording medium may include a hard disk, a floppy disk, a magnetic tape, an optical recording medium, a floptical disk, and a hardware device specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc.

[0225] FIG. 8 is a diagram showing the data structure of a transition record related to one embodiment of the present invention.

[0226] Referring to FIG. 8, the transition record (900) may include at least one of a rule identifier (910), a rule type (920), a rule content (930), an application subject (940), an application segment (950), a content format (960), an expiration condition (970), a promotion score (980), and a validation condition (990).

[0227] The fields illustrated in FIG. 8 are not essential for implementing the transition record (900), and the transition record (900) may have more or fewer fields than the fields listed above.

[0228] The transition record (900) may be in the form that the operation rule generated by the operation rule generation unit (120) is stored in the storage unit (850). The term "transition record" refers to a data structure in which the operation rule is stored, and includes both cases where the operation rule is a transition rule and a evasion rule. That is, "transition" simply means that a rule derived from a reaction result obtained from the first account (210) is transferred to the creation of content in the second account (220), and does not limit whether the rule is a rule used for the creation of the second content or a rule that restricts the creation of the second content. The transition record (900) may not include the original text of the first content, but may include the scope and conditions to which the operation rule is applied.

[0229] The rule identifier (910) may be information for identifying an operation rule. The transition record (900) may further include information for identifying the first content from which the operation rule was derived. Accordingly, it may be tracked whether a specific operation rule was derived from the response result of the first content.

[0230] The rule type (920) may be information indicating whether the operation rule is a transition rule or an avoidance rule. In one embodiment, the rule type (920) may further indicate whether the operation rule is a rule defining the subject of the second content, a rule defining the style, a rule defining the response method, a rule regarding response mode weighting, or a rule regarding time-based operation weighting.

[0231] Rule content (930) may be information indicating the content prescribed by the operating rules. Rule content (930) does not include the original text of the first content and may indicate at least one of the topic pattern, stylistic features, response-inducing elements, negative response elements, and avoidance conditions of the first content. For example, rule content (930) may include information that a question-type introduction is effective for content regarding technology comparison.

[0232] The applicable topic (940) may be information indicating the topic to which the operating rule applies. The applicable segment (950) may be information indicating the segment to which the operating rule applies, and said segment may be distinguished according to the follower's attributes, region, language, or interests. The content format (960) may be information indicating the format of the content to which the operating rule applies, said format may include posts, comments, or reaction displays.

[0233] In this way, by including the application subject (940), application segment (950), and content format (960) in the transition record (900), it is possible to prevent the indiscriminate application of an operation rule derived from a specific condition to the creation of a second content unrelated to that condition.

[0234] The expiration condition (970) may be information indicating a condition under which an operating rule is no longer applied. The expiration condition may include at least one of an expiration time, an upper limit on the number of applications, and a condition requiring re-verification. Accordingly, it is prevented that an operating rule that has become invalid over time continues to be applied.

[0235] The promotion score (980) may be information indicating the promotion score calculated for the first content from which the operation rule was derived. The transition record (900) may further include information indicating the sample size and policy risk level of the response result from which the operation rule was derived. Accordingly, the reliability of the operation rule can be managed together.

[0236] The verification condition (990) may be information indicating conditions that must be satisfied for the operation rule to be applied to the creation of the second content. The verification condition may include at least one of brand tone suitability, an upper limit on the degree of policy risk, and whether human approval is required. The verification condition may be used as a standard for verification performed by the rule verification unit (130).

[0237] In one embodiment, the transition record (900) may be stored in the form of structured data as follows.

[0238] { "knowledge_id": "k-204", "source_content": "c-77", "knowledge_type": "response_mode_weight", "abstract_rule": "question-style opening works better for technical comparison posts", "promotion_score": 0.81, "sample_size": 420, "policy_risk_score": 0.09, "required_gate": ["brand_tone", "low_policy_risk", "human_approval"]}

[0239] The storage unit (850) can store a transition record (900), and the content creation unit (140) can determine whether to apply the corresponding operation rule to the second content to be created by referring to at least one of the application topic (940), application segment (950), content format (960), and expiration condition (970) included in the stored transition record (900).

[0240] In one embodiment, when the operation rule is an avoidance rule, the rule type (920) indicates that it is an avoidance rule, the rule content (930) indicates a prohibited expression, a prohibited topic, or an upper limit on the intensity of the expression, and the verification condition (990) may indicate whether human approval is required.

[0241] As described above, according to one embodiment of the present invention, an operation rule is generated based on the reaction result to the first content posted through the first account (210), and the second content of the second account (220) is generated based on the operation rule without copying the first content, thereby allowing a content strategy to be explored while lowering the reputation risk and sanction risk of the second account (220). Additionally, the characteristics of the first content with a good reaction result are converted into a transition rule, and the characteristics of the first content with a poor reaction result or confirmed policy risk are converted into an avoidance rule, respectively; only the operation rule that has passed verification is applied to the generation of the second content, and the operation rule can be repeatedly updated by using the reaction result to the second content again for updating the operation rule.

[0242] Meanwhile, the method according to the present invention is written as a program executable on a computer and can be implemented on various recording media such as magnetic storage media, optical reading media, and digital storage media.

[0243] Implementations of the various technologies described herein may be implemented in digital electronic circuits, computer hardware, firmware, software, or a combination thereof. Such implementations may be implemented as computer programs stored on a computer-readable storage medium to control the operation of a data processing device. Such implementations may be implemented as computer program products, i.e., information carriers, such as computer programs stored on a machine-readable storage device or a computer-readable non-transient storage medium, for processing by the operation of a data processing device, e.g., a programmable processor, a computer, or a plurality of computers, or for controlling such operation. Computer programs such as the computer program(s) described above may be written in any form of programming language including compiled or interpreted languages, and may be deployed in any form including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. Computer programs may be deployed to be processed on one computer or a plurality of computers at a single site, or distributed across a plurality of sites and interconnected by a communication network.

[0244] Processors suitable for processing computer programs include, for example, both general-purpose and special-purpose microprocessors, and any one or more processors of any type of digital computer. Generally, the processor will receive instructions and data from read-only memory or random access memory, or both. The elements of the computer may include at least one processor that executes instructions and one or more memory devices that store instructions and data. Generally, the computer may include one or more mass storage devices that store data, for example, magnetic, magneto-optical disks, or optical disks, or may be combined to receive data from these, transmit data to these, or both. Information carriers suitable for embodying computer program instructions and data include, for example, semiconductor memory devices, magnetic media such as hard disks, floppy disks, and magnetic tapes; optical media such as CD-ROMs (Compact Disk Read Only Memory) and DVDs (Digital Video Disks); magneto-optical media such as floptical disks; ROMs (Read Only Memory); RAMs (Random Access Memory); flash memory; EPROMs (Erasable Programmable ROM); EEPROMs (Electrically Erasable Programmable ROM); etc. Processors and memory may be supplemented by or included in special-purpose logic circuit organizations.

[0245] Although this specification contains details of a number of specific embodiments, they should not be understood as limiting the scope of any invention or claimables, but rather as descriptions of features that may be characteristic of a specific embodiment of a specific invention. Specific features described in this specification in the context of individual embodiments may be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any appropriate sub-combination. Furthermore, while features may operate in a specific combination and be described as initially claimed, one or more features from the claimed combination may be excluded from the combination in some cases, and the claimed combination may be changed to a sub-combination or a variation of the sub-combination.

[0246] Likewise, although operations are depicted in the drawings in a specific order, this should not be understood as requiring that such operations be performed in that specific or sequential order depicted to obtain a desirable result, or that all depicted operations must be performed. In certain cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of the various device components of the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and devices can generally be integrated together into a single software product or packaged into multiple software products.

[0247] Meanwhile, the embodiments of the present invention disclosed in this specification and drawings are merely specific examples provided to aid understanding and are not intended to limit the scope of the present invention. It is obvious to those skilled in the art that other variations based on the technical concept of the present invention are possible in addition to the embodiments disclosed herein.

[0248] [Explanation of the symbol]

[0249] 100: Content creation device

[0250] 110: Reaction Result Acquisition Unit

[0251] 120: Operation Rule Creation Section

[0252] 121: Evaluation Department

[0253] 122: Classification section

[0254] 123: Transition rule generator

[0255] 124: Evasion rule generation section

[0256] 130: Rule Verification Unit

[0257] 140: Content Creation Section

[0258] 200: Platform

[0259] 210: 1st account

[0260] 220: Second account

[0261] 810: Processor

[0262] 820: Memory

[0263] 830: Communications Department

[0264] 840: Bus

[0265] 850: Storage section

[0266] 900: Transition Record

[0267] 910: Rule Identifier

[0268] 920: Rule Type

[0269] 930: Rule Details

[0270] 940: Application Topic

[0271] 950: Applied segment

[0272] 960: Content format

[0273] 970: Expiration condition

[0274] 980: Promotion points

[0275] 990: Verification conditions

Claims

Claim 1 A content creation device comprising: a reaction result acquisition unit for acquiring a reaction result for a first content posted through a first account; an operation rule generation unit for generating an operation rule based on the reaction result; and a content creation unit for generating a second content to be posted through a second account separated from the first account based on the operation rule without copying the first content, wherein the operation rule is a rule derived from the first content and stipulating a method of generating the second content. Claim 2 A content generating device according to claim 1, wherein the reaction result includes a quantitative reaction indicator and a policy risk signal for the first content, and the policy risk signal includes at least one of a report, deletion, exposure restriction, warning, and account status change. Claim 3 A content generation device according to claim 1, wherein the operation rule generation unit classifies the first content into either a transition target or an avoidance target based on the reaction result, generates a transition rule used for generating the second content when classified as a transition target, and generates an avoidance rule that restricts the generation of the second content when classified as an avoidance target. Claim 4 A content generating device according to paragraph 3, wherein the avoidance rule is derived from at least one of the negative reaction element and avoidance condition of the first content, and includes at least one of a prohibited expression, a prohibited topic, an expression intensity upper limit, and a prohibited expression proximity. Claim 5 A content generating device according to paragraph 3, wherein the transition rule is derived from at least one of the topic pattern, stylistic features, and response-inducing elements of the first content, and defines at least one of the topic, style, and response method of the second content. Claim 6 In claim 1, the content generating device further comprises a rule verification unit that verifies whether the operating rule satisfies a predetermined standard, and the content generating unit generates the second content based on the operating rule that has passed the verification. Claim 7 A content generation device according to claim 1, wherein the operation rule generation unit evaluates the response result using a quantitative response indicator for the first content and a natural language response for the first content together as inputs, and generates the operation rule based on the evaluation. Claim 8 A content generation device according to paragraph 3, wherein the operation rule generation unit calculates a promotion score by using the degree of reaction of the first content as an increasing factor and at least one of the degree of policy risk of the first content and the sample uncertainty of the reaction result as a decreasing factor, and determines whether to classify the first content as a transition target based on the promotion score. Claim 9 A content creation device according to claim 1, wherein the operating rule is stored as a transition record comprising at least one of an applicable subject, an applicable segment, a content format, and an expiration condition. Claim 10 A content creation device according to claim 1, wherein the reaction result acquisition unit further acquires a reaction result for the second content posted through the second account, and the operation rule generation unit further updates the operation rule based on the reaction result for the second content. Claim 11 A content creation method performed by a content creation device including a processor and memory, comprising: a step in which the content creation device obtains a reaction result for a first content posted through a first account; a step in which the content creation device generates an operation rule based on the reaction result; and a step in which the content creation device generates a second content to be posted through a second account separated from the first account based on the operation rule without copying the first content, wherein the operation rule is a rule derived from the first content and prescribes a method of generating the second content. Claim 12 A content creation method according to claim 11, wherein the response result includes a quantitative response indicator and a policy risk signal for the first content, and the policy risk signal includes at least one of a report, deletion, exposure restriction, warning, and account status change. Claim 13 A method for generating content according to claim 11, wherein the step of generating the operating rule comprises: a step of classifying the first content into either a transition target or an avoidance target based on the reaction result; a step of generating a transition rule used for generating the second content when classified as a transition target; and a step of generating an avoidance rule that restricts the generation of the second content when classified as an avoidance target. Claim 14 A content creation method according to claim 13, wherein the avoidance rule is derived from at least one of the negative reaction element and avoidance condition of the first content, and includes at least one of a prohibited expression, a prohibited topic, an expression intensity upper limit, and a proximity of a prohibited expression. Claim 15 A content generation method according to claim 13, wherein the transition rule is derived from at least one of the topic pattern, stylistic features, and response-inducing elements of the first content, and defines at least one of the topic, style, and response method of the second content. Claim 16 In claim 11, the content generation method further comprises the step of the content generation device verifying whether the operating rule satisfies a predetermined standard, and the step of generating the second content comprises the step of generating the second content based on the operating rule that passed the verification. Claim 17 A content generation method according to claim 11, wherein the step of generating the operating rule comprises evaluating the response result using a quantitative response indicator for the first content and a natural language response for the first content together as inputs, and generating the operating rule based on the evaluation. Claim 18 A method for generating content according to claim 13, wherein the step of generating the operating rule comprises: a step of calculating a promotion score by using the degree of reaction of the first content as an increasing factor and at least one of the degree of policy risk of the first content and the sample uncertainty of the reaction result as a decreasing factor; and a step of determining whether to classify the first content as a transition target based on the promotion score. Claim 19 In paragraph 11, the above-mentioned operating rule is a content creation method stored as a transition record comprising at least one of an applicable subject, an applicable segment, a content format, and an expiration condition. Claim 20 In claim 11, the content creation method further comprises the step of the content creation device further obtaining a reaction result for the second content posted through the second account; and the step of the content creation device further updating the operation rule based on the reaction result for the second content.

Citation Information

Patent Citations

  • Mobile memory system for secure storage and delivery of media content

    CN101361076A

  • The system which operates a on-line communication Account

    KR101657991B1

  • A Marketing information analysis service providing apparatus for performing content power index analysis of marketing channels using channel element data and prior index data, and method of operation thereof

    KR1020250098318A

  • Method and system for ai-based content evaluation

    KR1020250132345A