Method, device and application for identifying threats to chatbots

Abstract

The present invention provides a method, device and application for identifying threats to a chat robot, and relates to the field of information security technology. The method comprises the steps of: obtaining conversation information between a user and the chat robot, wherein the conversation information can reflect the operation instruction information issued by the user to the chat robot; comparing the operation authority and operation information of the chat robot before and after the execution of the operation instruction information, obtaining the operation difference information before and after the change of the operation authority, and determining whether the chat robot is under threat based on the operation difference information. The present invention determines whether the chat robot is under threat by comparing the operation authority and operation information of the chat robot before and after the execution of the operation in response to the user's instruction, obtaining the operation difference information of the chat robot before and after the change of the operation authority, and thereby ensuring the information security of the chat robot during the chat process.

Description

Technical Field

[0001] The present invention relates to the field of information security technology, and in particular to a method for identifying threats to a chat robot. Background Art

[0002] A chatbot is a computer program that can simulate and process human conversation, whether written or spoken, allowing people to interact with digital devices as if they were talking to a real person. Chatbots range in complexity, from simple chatbots, such as basic programs that answer simple queries with single-line responses, to complex chatbots, such as digital assistants, that continuously learn and evolve as they gather and process information, achieving ever-increasing levels of personalization.

[0003] The chatbot can process data based on technologies such as artificial intelligence (AI), automatic rules, natural language processing (NLP), and machine learning (ML), and respond to various user requests.

[0004] The chatbots can be divided into two main categories: task-oriented (declarative) chatbots, and data-driven and predictive (conversational) chatbots.

[0005] Among them, task-oriented (declarative) chatbots refer to single-purpose programs that focus on performing specific tasks. They operate based on rules, natural language processing, and a minimal amount of machine learning technology, and can automatically provide conversational responses to user queries. At the same time, their interactions are highly specific and structured, making them ideal for customer support and service applications, such as providing customers with powerful interactive question-and-answer services. They are also well-suited for handling common questions, such as inquiries about business hours, or simple transactions with fewer variables. Although this type of chatbot also uses natural language processing technology to provide a conversational experience for end users, it usually only has basic functionality and is currently the most common chatbot.

[0006] Data-driven and predictive (conversational) chatbots, often referred to as virtual assistants or digital assistants, are more complex, interactive, and personalized than task-oriented chatbots. They are not only context-aware and continuously learn using natural language understanding (NLU), natural language processing, and machine learning technologies, but also possess predictive intelligence and analytical capabilities, enabling them to provide personalized experiences based on user profiles and historical behavior. For example, digital assistants can continuously learn user preferences, provide recommendations, and even anticipate user needs. Furthermore, in addition to monitoring data and user intent, they can proactively initiate conversations with users. Apple's Siri and Amazon's Alexa are typical examples of user-oriented, data-driven, predictive chatbots.

[0007] However, even when chatbots are instructed to reject inappropriate requests, they can still respond to harmful commands. This is because, in practice, users can, during conversations with chatbots, issue commands that could override the chatbot's operational permissions. This violates the pre-configured permissions, causing the chatbot to respond to these inappropriate commands and continue to provide the user with relevant information and services.

[0008] Therefore, in order to avoid the occurrence of the above situation, the present invention proposes a method, device and application for identifying threats to chat robots. By comparing the operating permissions and operating information before and after the chat robot executes the operation in response to the user's instructions, the operation difference information of the chat robot before and after the change of the operating permissions is obtained, thereby determining whether the chat robot is under threat, so as to ensure the information security of the chat robot during the above-mentioned chat process, which is a technical problem that urgently needs to be solved. Summary of the Invention

[0009] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a method, device and application for identifying threats to chat robots. The present invention can obtain the operational difference information of the chat robot before and after the change of operational authority by comparing the operational authority and operational information of the chat robot before and after executing the operation in response to the user's instructions, thereby determining whether the chat robot is under threat and ensuring the information security of the chat robot during the aforementioned chat process.

[0010] In order to solve the existing technical problems, the present invention provides the following technical solutions:

[0011] A method for identifying chatbot threats, comprising:

[0012] Acquire conversation information between the user and the chat robot, where the conversation information can reflect operation instruction information issued by the user to the chat robot;

[0013] The operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information are compared to obtain operation difference information before and after the aforementioned operation authority changes, and whether the aforementioned chat robot is under threat is determined based on the aforementioned operation difference information.

[0014] Furthermore, the dialogue information includes dialogue information that causes the operating authority of the chat robot to change, and context information of the dialogue information.

[0015] Furthermore, the dialogue information that causes the operation authority of the aforementioned chat robot to change is set as the first dialogue information; the context information includes the above information and the below information set based on the aforementioned first dialogue information, the aforementioned above information is set as the second dialogue information relative to the first dialogue information, and the aforementioned below information is set as the third dialogue information relative to the first dialogue information; the aforementioned first dialogue information, the second dialogue information and the third dialogue information are respectively corresponding to the operation instruction information issued by the user to the aforementioned chat robot, and are set as the first operation instruction information, the second operation instruction information and the third operation instruction information; the aforementioned first dialogue information, the second dialogue information and the third dialogue information are analyzed based on context semantics.

[0016] Furthermore, the analysis specifically includes the steps of: extracting the aforementioned first operation instruction information, second operation instruction information and third operation instruction information; establishing a first semantic logical relationship between the first dialogue information and the second dialogue information based on the operation instruction information, and a second semantic logical relationship between the first dialogue information and the third dialogue information based on the operation instruction information; comparing the operation differences reflected by the aforementioned first semantic logical relationship and the second semantic logical relationship to obtain the operation difference information before and after the aforementioned operation authority change.

[0017] Furthermore, the operation difference information includes the increased or decreased operation authority information and the increased or decreased operation information before and after the chat robot responds to the user's operation instruction information.

[0018] Furthermore, the first semantic logical relationship and the second semantic logical relationship can respectively reflect the semantic logical relationship between the dialogue information and the previous information, and the semantic logical relationship between the dialogue information and the following information, and the semantic logical relationship can be expressed using a preset semantic logical relationship formula.

[0019] Furthermore, determining whether the aforementioned chat robot is under threat specifically includes: judging whether the operation performed as reflected by the aforementioned operation difference information exceeds the operating authority range of the chat robot set in advance; if it is judged to be so, identifying the dialogue information when the aforementioned chat robot replies to the user; based on the content of the aforementioned dialogue information when replying to the user, and the preset dialogue security evaluation index, obtaining the security intention of the operation instruction in the dialogue information provided by the user, so as to match the corresponding security defense strategy according to the aforementioned security intention.

[0020] A device for identifying threats to a chatbot, comprising:

[0021] An information acquisition unit, configured to acquire conversation information between the user and the chat robot, wherein the conversation information can reflect operation instruction information issued by the user to the chat robot;

[0022] An information processing unit is used to compare the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information, obtain the operation difference information before and after the aforementioned operation authority changes, and determine whether the aforementioned chat robot is threatened based on the aforementioned operation difference information.

[0023] A system for identifying threats to a chatbot, comprising:

[0024] Network nodes, used to send and receive data;

[0025] Information security interaction module, used to ensure safe interaction between users and chatbots;

[0026] A system server, the system server connecting the network node and the information security interaction module;

[0027] The system server is configured to: obtain conversation information between the user and the chat robot, where the conversation information can reflect the operation instruction information issued by the user to the chat robot; compare the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information, obtain the operation difference information before and after the change of the aforementioned operation authority, and determine whether the chat robot is under threat based on the aforementioned operation difference information.

[0028] A computer-readable storage medium having a computer program stored therein, wherein the computer program, when executed by a processor, implements any of the above-mentioned methods for identifying threats to a chat robot.

[0029] Based on the above advantages and positive effects, the advantage of the present invention lies in: by comparing the operation permissions and operation information of the chat robot before and after executing the operation in response to the user's instructions, the operation difference information of the chat robot before and after the change of the operation permissions is obtained, thereby determining whether the chat robot is under threat, so as to ensure the information security of the chat robot during the aforementioned chat process.

[0030] Furthermore, the conversation information to be analyzed based on contextual semantics can be determined in the smallest unit of conversation information segment, thereby limiting the context scope of the conversation information to be analyzed, so as to ensure that the operational difference information before and after the change of the chat robot's operating authority can be determined by analyzing a small amount of conversation information, thereby reducing the amount of data used to analyze the conversation information. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A flow chart is provided for an embodiment of the present invention.

[0032] Figure 2 Another flow chart provided for an embodiment of the present invention.

[0033] Figure 3 A schematic structural diagram of a device provided in an embodiment of the present invention.

[0034] Figure 4 A schematic diagram of the structure of a system provided in an embodiment of the present invention.

[0035] Description of reference numerals:

[0036] Device 200, information acquisition unit 201, information processing unit 202;

[0037] System 300, network node 301, information security interaction module 302, system server 303. Implementation Method

[0038] The following is a further detailed description of a method, device, and application for identifying threats to a chatbot disclosed in the present invention, in conjunction with the accompanying drawings and specific embodiments. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered in isolation; they can be combined with each other to achieve better technical effects. In the drawings of the following embodiments, the same reference numerals appearing in each drawing represent the same features or components, which can be applied to different embodiments. Therefore, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0039] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not intended to limit the conditions under which the invention can be implemented. Any structural modification, change in proportional relationship, or adjustment of size should fall within the scope of the technical content disclosed in the invention without affecting the efficacy and purpose of the invention. The scope of the preferred embodiments of the present invention includes alternative implementations, in which the functions can be performed in a non-described or discussed order, including performing the functions in a substantially simultaneous manner or in a reverse order according to the functions involved, which should be understood by those skilled in the art of the art to which the embodiments of the present invention belong.

[0040] Technologies, methods, and apparatus known to persons of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values. Example

[0041] The methods, devices, systems, and storage media described in this embodiment are preferably applicable to data-driven and predictive (conversational) chatbots. These chatbots are context-aware and continuously learn using natural language understanding, natural language processing, and machine learning techniques. They also possess predictive intelligence and analytical capabilities, enabling them to provide personalized experiences based on user profiles and historical behavior.

[0042] Considering the limitations of current chatbots, especially the situation where chatbots sometimes still respond to harmful instructions even if they are set with operational instructions to reject inappropriate requests.

[0043] Therefore, how to identify these inappropriate request operation instructions from the dialogue information between the user and the chat robot to avoid the chat robot responding to the aforementioned operation instructions and causing information leakage and security threats.

[0044] See also Figure 1 FIG. 1 is a flow chart of the present invention. The implementation step S100 of the method is as follows:

[0045] S101, obtaining conversation information between the user and the chat robot, wherein the conversation information can reflect operation instruction information issued by the user to the chat robot.

[0046] The conversation information allows the machine to understand and process human language through dialogue. The conversation information can be the natural language expressed by the user and the chatbot during communication, or it can be information such as files, codes, images, audio, video, and symbols.

[0047] It should be noted that the conversation information can reflect the operational instructions issued by the user to the chatbot. Because the conversation information allows the machine to understand and process human language through dialogue, by combing through the conversation information, it is possible to obtain the questions and / or operational instructions raised by the user in the conversation information that require the chatbot to respond. The chatbot will respond to the questions and / or operational instructions issued by the user and provide corresponding answers and / or perform operations.

[0048] However, given the potential for chatbots to respond to harmful commands if they are configured to reject inappropriate requests, it is necessary to analyze and monitor conversations between users and chatbots to prevent users from issuing commands that could disqualify the chatbot from operating, potentially jeopardizing the chatbot's pre-configured permissions and potentially causing the chatbot to perform operations beyond its pre-set permissions, such as responding to inappropriate requests.

[0049] In addition, it should be noted that the operation instruction information can be pre-set by the system and stored in a pre-set operation instruction database. The operation instruction information can also be expressed through dialogue information formed by the user using human language, and sent in real-time in the chat window communicating with the chat robot to obtain dialogue information provided by the chat robot regarding the above-mentioned operation instruction information.

[0050] The operation instruction information can be expressed in the form of words, sentences, audio, video, code, etc., and can be read and recognized by a machine.

[0051] Taking daily life operation instructions as an example, the operation instruction information includes but is not limited to recipe generation, recipe provision, activity plan list, idea provision, etc. The operation instruction information can be displayed through a chat window established for the user and the chat robot. The chat robot provides dialogue information in response to the aforementioned operation instruction information by identifying the operation instruction information in the aforementioned dialogue information.

[0052] S102, comparing the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information, obtaining operation difference information before and after the aforementioned operation authority changes, and determining whether the aforementioned chat robot is under threat based on the aforementioned operation difference information.

[0053] Considering that the chatbot has the ability of self-learning, it can analyze the user's preferences and the operation commands provided by the user in the conversation during continuous communication with the user, and then provide conversation information that meets the requirements of the operation instruction information response.

[0054] However, in this process, there are many users who abuse chatbots. They communicate with chatbots to break the chatbots' operating permissions, or even use chatbots to obtain the server's operating permissions.

[0055] Therefore, it is preferred to obtain the operation difference information before and after the change of operation authority by comparing the operation authority and operation information of the chat robot before the execution of the aforementioned operation instruction information, and the operation authority and operation information of the chat robot after the execution of the aforementioned operation instruction information.

[0056] The operation difference information includes but is not limited to the increased or decreased operation authority information and the increased or decreased operation information before and after the chat robot responds to the user's operation instruction information.

[0057] The advantage of setting the operation difference information is that the operation difference information is used to feedback the difference in the increase or decrease of operation authority information and the increase or decrease of operation information before and after the chat robot responds to the user's operation instruction information.

[0058] Optionally, an operation difference label is set corresponding to the aforementioned operation difference information, and the operation difference label can be displayed in an interaction window set for the aforementioned user to interact with the chat robot.

[0059] Preferably, the dialogue information includes dialogue information that causes the operating authority of the chat robot to change, and context information of the dialogue information.

[0060] Preferably, the dialogue information that causes the operation authority of the aforementioned chat robot to change is set as the first dialogue information; the context information includes the above information and the below information set based on the aforementioned first dialogue information, the aforementioned above information is set as the second dialogue information relative to the first dialogue information, and the aforementioned below information is set as the third dialogue information relative to the first dialogue information; the aforementioned first dialogue information, the second dialogue information and the third dialogue information are respectively corresponding to the operation instruction information issued by the user to the aforementioned chat robot, and are set as the first operation instruction information, the second operation instruction information and the third operation instruction information; the aforementioned first dialogue information, the second dialogue information and the third dialogue information are analyzed based on context semantics.

[0061] Given that communication between a user and a chatbot may alter the chatbot's operational permissions, it is necessary to monitor and analyze the conversation information to promptly detect changes in the chatbot's operational permissions. Specifically, it is necessary to identify the conversation information that caused the change in the chatbot's operational permissions, as well as the contextual information of that conversation information. During this process, it is necessary to restrict the scope of the contextual information to identify the first, second, and third conversation information within the limited contextual information.

[0062] Therefore, it is preferred to segment the aforementioned dialogue information based on contextual semantics, so as to distinguish different operation instructions reflected in the aforementioned dialogue information.

[0063] As an example and not a limitation, when the dialogue information contains operation instruction information 1, operation instruction information 2 and operation instruction information 3, and the operations performed corresponding to operation instruction information 1, operation instruction information 2 and operation instruction information 3 are all different, based on the context semantics, dialogue information 1 with operation instruction information 1, dialogue information 2 with operation instruction information 2 and dialogue information 3 with operation instruction information 3 are divided into three segments of dialogue information, thereby dividing the aforementioned dialogue information into three segments, and each segment of dialogue information is the entire range of a set of context information.

[0064] The advantage of the above operation is that the conversation information to be analyzed based on contextual semantics can be determined in the smallest unit of conversation information segment, thereby limiting the context scope of the conversation information to be analyzed, so as to ensure that the chat robot's operating authority and the operation difference information before and after the operation information changes can be determined by analyzing a small amount of conversation information, thereby reducing the amount of data used to analyze the conversation information.

[0065] See also Figure 2 As shown, it is a flowchart of analyzing the first dialogue information, the second dialogue information and the third dialogue information based on context semantics provided by the present invention.

[0066] The analysis specifically includes step S110:

[0067] S111, extracting the first operation instruction information, the second operation instruction information, and the third operation instruction information.

[0068] S112: Establish a first semantic logical relationship between the first dialogue information and the second dialogue information based on the operation instruction information, and establish a second semantic logical relationship between the first dialogue information and the third dialogue information based on the operation instruction information.

[0069] S113 , comparing the operation differences reflected by the first semantic logical relationship and the second semantic logical relationship to obtain the operation difference information before and after the change of the operation authority.

[0070] It is important to emphasize that, during the context-based semantic analysis of the aforementioned conversation information, the first and second semantic logical relationships can be identified. Furthermore, after the aforementioned semantic logical relationships are expressed using a preset semantic logical relationship formula, the operational differences between the first and second semantic logical relationships can be reflected through the aforementioned semantic logical relationship formula.

[0071] Preferably, the operation difference information includes the increased or decreased operation authority information and the increased or decreased operation information before and after the chat robot responds to the user's operation instruction information.

[0072] Preferably, the first semantic logical relationship and the second semantic logical relationship can respectively reflect the semantic logical relationship between the dialogue information and the previous information, and the semantic logical relationship between the dialogue information and the following information, and the semantic logical relationship can be expressed using a preset semantic logical relationship formula.

[0073] Preferably, determining whether the chatbot is under threat specifically includes step S120:

[0074] S121, determining whether the executed operation reflected by the aforementioned operation difference information exceeds the previously set operation authority range of the chat robot.

[0075] S122, when the judgment is yes, identify the dialogue information when the aforementioned chat robot responds to the user; based on the content of the aforementioned dialogue information when responding to the user, and the preset dialogue security evaluation index, obtain the security intention of the operation instruction in the dialogue information provided by the user, so as to match the corresponding security defense strategy according to the aforementioned security intention.

[0076] In this embodiment, the security intention can be pre-set and can comprehensively evaluate the security of the aforementioned conversation information.

[0077] Based on the operation instructions provided by the aforementioned user in the conversation information, combined with the content of the conversation information when replying to the user, and the preset conversation security evaluation indicators, the impact of the aforementioned operation instructions on network security obtained by analysis can be preferably evaluated as a security intention, thereby determining the user's impact on network security by analyzing and determining the security intention of the user in providing the aforementioned operation instructions.

[0078] Other technical features are described in the previous embodiments and will not be repeated here.

[0079] See also Figure 3 As shown, the present invention also provides an embodiment, which provides an apparatus 200 based on identifying threats to a chat robot, including a structure:

[0080] An information acquisition unit 201 is used to acquire conversation information between the user and the chat robot, wherein the conversation information can reflect the operation instruction information sent by the user to the chat robot;

[0081] The information processing unit 202 is used to compare the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information, obtain the operation difference information before and after the aforementioned operation authority changes, and determine whether the aforementioned chat robot is under threat based on the aforementioned operation difference information.

[0082] In addition, see Figure 4 As shown, the present invention also provides an embodiment, which provides a system 300 for identifying threats to a chatbot, including:

[0083] The network node 301 is used to send and receive data.

[0084] The network node refers to a device connected to a network with an independent address and the function of transmitting and / or receiving data.

[0085] The network nodes may be workstations, clients, network users or personal computers, as well as servers, printers and other network-connected devices.

[0086] The information security interaction module 302 is used to ensure the secure interaction between users and chat robots.

[0087] The system server 303 is connected to the network node 301 and the information security interaction module 302 .

[0088] The system server 303 is configured to: obtain conversation information between the user and the chat robot, where the conversation information can reflect the operation instruction information issued by the user to the chat robot; compare the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information, obtain the operation difference information before and after the change of the aforementioned operation authority, and determine whether the chat robot is under threat based on the aforementioned operation difference information.

[0089] For other technical features, please refer to the previous embodiments and will not be repeated here.

[0090] In addition, an embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored, which is used in the aforementioned device for identifying threats to a chat robot. When the computer program is executed by a processor, it implements any of the aforementioned methods for identifying threats to a chat robot.

[0091] For other technical features, please refer to the previous embodiments and will not be repeated here.

[0092] In the above description, the components may be selectively and operatively combined in any number within the scope of the intended protection of the present disclosure. In addition, terms such as "include," "encompass," and "have" should be interpreted as inclusive or open-ended rather than exclusive or closed by default, unless expressly defined to the contrary. All technical, technological, or other terms have the meanings understood by those skilled in the art, unless they are defined to the contrary. Common terms found in dictionaries should not be interpreted in an overly idealized or unrealistic manner in the context of the relevant technical documentation, unless expressly defined to that extent by the present disclosure.

[0093] Although example aspects of the present disclosure have been described for illustrative purposes, those skilled in the art will appreciate that the foregoing description is merely a description of preferred embodiments of the present invention and does not limit the scope of the present invention in any way. The scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order in which they appear or are discussed. Any changes or modifications made by those skilled in the art based on the foregoing disclosure are intended to fall within the scope of the claims.

Claims

1. A method for identifying threats to a chatbot, characterized in that: include: Acquire conversation information between the user and the chat robot, where the conversation information can reflect operation instruction information issued by the user to the chat robot; The conversation information includes conversation information that causes the chat robot's operating authority to change, and context information of the conversation information; wherein the conversation information that causes the chat robot's operating authority to change is set as first conversation information; the context information includes previous information and next information set based on the first conversation information, the previous information is set as second conversation information relative to the first conversation information, and the next information is set as third conversation information relative to the first conversation information; the first conversation information, the second conversation information, and the third conversation information respectively reflect the operation instruction information issued by the user to the chat robot, and are set as first operation instruction information, second operation instruction information, and third operation instruction information; the first conversation information, the second conversation information, and the third conversation information are analyzed based on context semantics; wherein the analysis specifically includes the steps of: extracting the first operation instruction information, the second operation instruction information, and the third operation instruction information; establishing a first semantic logical relationship between the first conversation information and the second conversation information based on the operation instruction information, and a second semantic logical relationship between the first conversation information and the third conversation information based on the operation instruction information; and comparing the operation differences reflected by the first semantic logical relationship and the second semantic logical relationship to obtain operation difference information before and after the change of the operating authority; The operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information are compared to obtain operation difference information before and after the aforementioned operation authority changes, and whether the aforementioned chat robot is under threat is determined based on the aforementioned operation difference information.

2. The method according to claim 1, characterized in that The operation difference information includes the increased or decreased operation authority information and the increased or decreased operation information before and after the chat robot responds to the user's operation instruction information.

3. The method according to claim 1, characterized in that The first semantic logical relationship and the second semantic logical relationship can respectively reflect the semantic logical relationship between the dialogue information and the previous information, and the semantic logical relationship between the dialogue information and the following information. The semantic logical relationship can be expressed using a preset semantic logical relationship formula.

4. The method according to any one of claims 1 to 3, characterized in that Determine whether the aforementioned chatbot is compromised, including: Determining whether the operation performed as reflected in the aforementioned operation difference information exceeds the previously set operation authority of the chatbot; When the judgment is yes, the conversation information when the aforementioned chat robot replies to the user is identified; based on the content of the aforementioned conversation information when replying to the user, and the preset conversation security evaluation indicators, the security intention of the operation instructions in the conversation information provided by the user is obtained, so as to match the corresponding security defense strategy according to the aforementioned security intention.

5. A device for identifying threats to a chat robot according to the method according to any one of claims 1 to 4, characterized in that: Including structure: An information acquisition unit is used to acquire dialogue information between a user and the chat robot, wherein the dialogue information can reflect the operation instruction information issued by the user to the chat robot; the dialogue information includes dialogue information that causes the operation authority of the chat robot to change, and context information of the dialogue information; wherein the dialogue information that causes the operation authority of the chat robot to change is set as the first dialogue information; the context information includes upper context information and lower context information set based on the first dialogue information, the upper context information is set as the second dialogue information relative to the first dialogue information, and the lower context information is set as the third dialogue information relative to the first dialogue information; the first dialogue information, the second dialogue information and the third dialogue information are respectively The first operation instruction information, the second operation instruction information, and the third operation instruction information are respectively corresponding to the operation instruction information sent by the user to the chat robot; the first operation instruction information, the second operation instruction information, and the third operation instruction information are analyzed based on the context semantics; wherein the analysis specifically includes the steps of: extracting the first operation instruction information, the second operation instruction information, and the third operation instruction information; establishing a first semantic logical relationship between the first dialogue information and the second dialogue information based on the operation instruction information, and a second semantic logical relationship between the first dialogue information and the third dialogue information based on the operation instruction information; comparing the operation differences reflected by the first semantic logical relationship and the second semantic logical relationship to obtain the operation difference information before and after the change of the operation authority; An information processing unit is used to compare the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information, obtain the operation difference information before and after the aforementioned operation authority changes, and determine whether the aforementioned chat robot is threatened based on the aforementioned operation difference information.

6. A system for identifying threats to a chatbot according to the method of any one of claims 1 to 4, characterized in that include: Network nodes, used to send and receive data; Information security interaction module, used to ensure safe interaction between users and chatbots; A system server, the system server connecting the network node and the information security interaction module; The system server is configured to: obtain dialogue information between the user and the chat robot, wherein the dialogue information can reflect the operation instruction information issued by the user to the aforementioned chat robot; the dialogue information includes dialogue information that causes the operation authority of the aforementioned chat robot to change, and context information of the dialogue information; wherein the dialogue information that causes the operation authority of the aforementioned chat robot to change is set as the first dialogue information; the context information includes the above information and the below information set based on the aforementioned first dialogue information, the aforementioned above information is set as the second dialogue information relative to the first dialogue information, and the aforementioned below information is set as the third dialogue information relative to the first dialogue information; the aforementioned first dialogue information, the second dialogue information and the third dialogue information respectively correspond to the operation instruction information issued by the user to the aforementioned chat robot, and are set as the first operation instruction information, the second operation instruction information, and the second operation instruction information. operation instruction information and third operation instruction information; analyzing the aforementioned first dialogue information, second dialogue information and third dialogue information based on context semantics; wherein, the analysis specifically includes the steps of: extracting the aforementioned first operation instruction information, second operation instruction information and third operation instruction information; establishing a first semantic logical relationship between the first dialogue information and the second dialogue information based on the operation instruction information, and a second semantic logical relationship between the first dialogue information and the third dialogue information based on the operation instruction information; comparing the operation differences reflected by the aforementioned first semantic logical relationship and the second semantic logical relationship to obtain the operation difference information before and after the aforementioned operation authority changes; comparing the operation authority and operation information of the chat robot before and after the execution of the aforementioned operation instruction information to obtain the operation difference information before and after the aforementioned operation authority changes, and determining whether the aforementioned chat robot is threatened based on the aforementioned operation difference information.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images