Method for analyzing continuous use willingness of user by artificial intelligence agent

Abstract

The invention relates to the technical field of data processing, in particular to a method for analyzing continuous use willingness of a user by an artificial intelligence agent. The perception of the user on the artificial intelligence agent is divided into functional perception, emotion perception and invasive perception; the method specifically comprises the following steps: artificial intelligence agency intelligence and continuous use willingness of a user; a psychological perception theory and a fair theory; the user perceives the willingness to use the artificial intelligence agent. According to the invention, the influence of the artificial intelligence agent on the continuous use intention of the user is explored from the intelligent perspective; from two perspectives of autonomy and learnability, discovering that the intelligent attribute of the artificial intelligence agency has a difference on the influence mechanism of the continuous use intention of the user under the action of an intermediary agent or not; three basic dimensions of functional perception, emotion perception and invasive perception are utilized to comprehensively research an influence mechanism of an artificial intelligence agent on continuous use willingness of a user from positive and negative angles.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for analyzing user's continued use intention by an artificial intelligence agent. Background Art

[0002] With the continuous development of data and information, big data, deep learning, the Internet of Things and other related technologies have optimized the service model of artificial intelligence and accelerated the integration of artificial intelligence and applications. The emergence of these technologies has improved the intelligence level of artificial intelligence, allowing human-computer interaction to shift from traditional graphic interfaces to interactive pages with natural language and emotional understanding. For example, XiaoIce, Xiaodu, Siri, Tmall Genie, etc., they can help users manage a large number of applications or devices, improve work efficiency, and provide companionship and entertainment. With the emergence of ChatGPT, the intelligence level of artificial intelligence agents has been further improved, and they can understand and execute more complex commands and tasks with their powerful intelligence. It combines reinforcement learning from user feedback, generates various types of information based on user-defined elements and styles, and continuously improves, demonstrating its intelligent advantages.

[0003] AI agents can also operate autonomously after user authorization and make decisions or actions based on user preferences. For example, Xiao Ai can automatically turn on the lights or air conditioning when the owner arrives home, and the intelligent agent can control the car to drive automatically. However, the intelligence of AI agents is not only reflected in their functional advantages. Now they have been improved to the point where they can understand user emotions and give positive emotional feedback, such as imitating human voice to communicate and provide care, analyzing user emotions and providing suggestions, etc. As the intelligence level of AI agents continues to improve, more users are paying more and more attention to interacting with AI agents. Compared with traditional self-service technologies, AI agents stand out for their excellent learning and adaptability. Their increasing intelligence enables them to serve users more flexibly and individually, increasing their satisfaction.

[0004] In fact, AI agents rely on huge databases to operate, especially the data and information entered by users. In order to better serve users, AI agents need to continuously collect user data, which may make users feel intrusive. For example, ChatGPT may provide hackers, cybercriminals, and intruders with more opportunities to infringe on privacy and rights; the personalized experience provided by AI agents may lead to privacy leakage; the autonomy of AI agents may limit the autonomy of users to some extent, limiting their experience. Obviously, the intelligence of AI agents can bring functional and emotional benefits to users. But on the other hand, a high level of intelligence may also make users feel the existence of risks and increase their concerns about security issues. Previous studies in the field of consumer behavior have mostly focused on initial adoption intention, but there are still studies that show that continued use intention is more explanatory, directly related to user expectations and satisfaction, and better reflect users' attitudes towards products. Therefore, the mechanism by which the intelligence level of AI agents affects users' continued use intention still needs further exploration.

[0005] The use of artificial intelligence agents is becoming more and more common, so it is necessary to conduct in-depth research on how the characteristics of artificial intelligence agents promote user perception and thus affect their continued use intention. Most of the previous studies on artificial intelligence agents started from the perspective of the impact of anthropomorphism, personalization, humor, service failure, etc. on user willingness to use. However, with the improvement of the intelligence level of artificial intelligence agents, there are fewer studies on the intelligence of artificial intelligence agents. Some literature on the intelligence of artificial intelligence agents is limited to exploring the advantages of the intelligence of artificial intelligence agents, but ignores that intelligence may also negatively affect user attitudes or behaviors; there are also studies that only discuss the negative impact of artificial intelligence agents on user behavior from a unilateral perspective, and there are fewer studies that comprehensively consider the positive and negative effects of the intelligence of artificial intelligence agents on user behavior. Summary of the invention

[0006] In order to make up for the shortcomings of the prior art, the present invention proposes an analysis method for an artificial intelligence agent on the user's continued use intention, and explores the impact of artificial intelligence agents on users' continued use intention from the perspective of intelligence; from the two perspectives of autonomy and learning, it is found that there are differences in the influence mechanism of the intelligent attributes of artificial intelligence agents on users' continued use intention with and without mediation; based on psychological perception theory and equity theory, the three basic dimensions of functional perception, emotional perception and intrusive perception are used to comprehensively study the influence mechanism of artificial intelligence agents on users' continued use intention from both positive and negative perspectives.

[0007] The technical solution adopted by the present invention to solve the technical problem is: the method for analyzing the user's continued use intention by an artificial intelligence agent described in the present invention divides the user's perception of the artificial intelligence agent into functional perception, emotional perception and intrusive perception; and analyzes the user's continued use intention from the following perspectives:

[0008] a. The intelligence of AI agents and the user's continued willingness to use them play a vital role in the user's acceptance and continued use of such products;

[0009] b. Theory of Mind and Equity Theory; Users evaluate the value of AI agents based on their social interactivity and personality traits; When users interact with AI agents, they tend to perceive them as human-like behavior;

[0010] c User perception and intention to use AI agents; the social support perceived by users during their interactions with non-human entities will ultimately affect their subsequent behavioral intentions; functional perception is closely related to personal services, which is reflected in completing commands or tasks issued by users and solving problems faced by users.

[0011] Preferably, a includes autonomy and willingness to continue using and learning and willingness to use; specifically as follows:

[0012] H1a. The autonomy of AI agents positively affects users’ continued use intention;

[0013] H1b. The learning ability of artificial intelligence agents positively affects users’ continued use intention.

[0014] Preferably, b includes autonomy and user perception and learning and consumer perception; specifically as follows:

[0015] H2a. The autonomy of AI agents positively affects users’ functional perception;

[0016] H2b. The autonomy of artificial intelligence agents positively affects consumers’ emotional perception;

[0017] H2c. The autonomy of AI agents positively affects consumers’ perception of intrusiveness;

[0018] H3a. The learning ability of AI agents positively affects users’ functional perception;

[0019] H3b. The learning ability of artificial intelligence agents positively affects users’ emotional perception;

[0020] H3c. The learning ability of artificial intelligence agents positively affects users’ perception of intrusiveness.

[0021] Preferably, c includes a psychological or emotional value enhancement effect that significantly enhances the user's continued use intention; specifically as follows:

[0022] H4a. When interacting with an AI agent, users’ functional perception has a positive impact on their continued use intention;

[0023] H4b. When interacting with an AI agent, users’ emotional perception has a positive impact on their continued use intention;

[0024] H4c. When interacting with AI agents, users’ perceived intrusiveness has a negative impact on their continued use intention.

[0025] The beneficial effects of the present invention are as follows:

[0026] First, users' functional perception, emotional perception, and intrusive perception of humanoid entities can significantly affect their continued use intention. We should attach importance to the key role of user perception, reasonably control the autonomy and learning level of artificial intelligence agents, stimulate users' functional and emotional perception, and make users perceive that the positive feedback on functions and emotions is higher than the negative impact of privacy invasion, thereby enhancing users' continued use intention.

[0027] Secondly, a high level of autonomy and learning can significantly enhance user functional and emotional perception to promote continued use intention; and prompt more careful improvement of the functions and features of AI agent products. For example, actively perceive changes in the environment (including physical environment and network environment) and autonomously solve problems that may arise for users; add human-like functions, including tone, sound, and emotional imitation, so that users are more likely to regard them as friends; consider the language differences of users, set dialect recognition or communicate in dialect, etc. When providing a variety of functional options, it is also necessary to focus on stimulating the emotional connection and satisfaction between users and AI agents during the interaction process.

[0028] Third, the improvement of autonomy will also increase the perception of intrusion. We should not blindly improve the level of product autonomy, but try to reduce the user's perception of intrusion during the interaction process. Improve the personalized autonomy function of artificial intelligence agents, for example, strengthen the two-way communication ability between AIA and users, give users selection and customization options; enhance the transparency and explainability of the AIA model decision-making process; avoid interference with the user's financial interests. [ Through these measures, users' sense of control and trust in the system are enhanced, user privacy is protected, and user trust is improved. At the same time, system security is regularly checked and publicly displayed to users to maintain stable user relationships.

[0029] In addition, a high level of learning will not increase the perception of intrusion. By improving the learning of AIA, the user's negative perception can be effectively reduced, resources can be rationally allocated, and the optimal combination of autonomy and learning can be achieved. By enhancing the learning of the system, existing or new information can be effectively used to provide users with more value, show higher benefits, solve more problems, and further enhance users' willingness to continue using. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The present invention is further described below in conjunction with the accompanying drawings and implementation modes.

[0031] Figure 1 It is a model diagram of the present invention;

[0032] Figure 2 It is a diagram of the model test results of the present invention. DETAILED DESCRIPTION

[0033] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further described below in conjunction with the accompanying drawings and preferred embodiments.

[0034] like Figure 1 As shown in the figure, a method for analyzing the user's continued use intention of an artificial intelligence agent is used to divide the user's perception of the artificial intelligence agent into functional perception, emotional perception, and intrusive perception; the user's continued use intention is analyzed from the following perspectives:

[0035] 1. The intelligence of AI agents and users’ continued willingness to use them play a vital role in users’ acceptance and continued use of such products;

[0036] 1.1 Autonomy and willingness to continue using

[0037] With the continuous upgrading of artificial intelligence technology, AIA's intelligence level is getting higher and higher, and it is becoming more and more autonomous. Autonomy, as one of the core attributes of AIA's intelligence, can perceive changes in the environment and actively provide corresponding suggestions or decisions. It is reflected in the autonomy of perception, thinking and action in the process of human-computer interaction, that is, autonomously perceive the environment in each task, think about the plan according to the environment, and take action according to the plan. For example, AIA can perceive changes in the user's emotions and actively turn on the user's favorite music.

[0038] Highly autonomous AI agents can better understand, predict and meet their needs, improving the overall user experience. When AI agents have higher autonomy, that is, they show stronger independent operation and self-adjustment capabilities when performing tasks, providing services or making decisions, they can better meet users' work or daily needs, thereby generating a higher degree of recognition. This sense of recognition stems from users' recognition of the advantages of autonomy, such as efficiency, personalized services and labor cost savings. User recognition will provide them with motivation to continue using AI agents.

[0039] Therefore, we hypothesize the following: H1a. The autonomy of AI agents positively affects users’ continued usage intention.

[0040] 1.2 Learning and willingness to use

[0041] Learning is the core competitiveness of AI agents. AI agents with high learning ability can continuously absorb and learn user habits and preferences through deep learning, reinforcement learning, etc., to achieve personalized services and improve user experience. AI agents rely on huge databases to provide services to users, and data is constantly updated, so its learning ability determines whether it can provide users with better services. The information support brought by the powerful learning ability of AI agents reduces the user's search cost and information asymmetry, improves decision-making efficiency, and will improve the user experience. A good user experience enhances the user's willingness to continue using.

[0042] Therefore, we hypothesize the following: H1b. The learnability of AI agents positively affects users’ continued use intention.

[0043] 2. Psychological Perception Theory and Equity Theory: Users evaluate the value of AI agents based on their social interactivity and personality traits; When users interact with AI agents, they tend to perceive them as human-like behaviors; Analyzing the positive or negative impact of AI attributes on user perception from a single perspective, few studies can fully explore the impact of AIA's intelligence on user perception from multiple dimensions and positive and negative perspectives. Based on psychological perception theory and equity theory, it is believed that the autonomy and learning of AIA may affect users' continued use behavior from three dimensions: function, emotion, and intrusiveness.

[0044] 2.1 Autonomy and User Perception

[0045] The autonomy of AI-related products is achieved through complex algorithms, which can complete the given tasks without or with little human intervention. The autonomy of AIA can be achieved by actively sensing and identifying the environment, actively providing suggestions or decisions, and increasing information support for users, for example, actively monitoring different needs of users through voice recognition. This kind of information support can make users tend to believe that AI algorithms are more accurate and reliable than human decisions, and perceive their stronger functionality. Therefore, the following hypothesis is proposed:

[0046] H2a. The autonomy of AI agents positively affects users’ functional perception.

[0047] The autonomy of AI agents can proactively provide suggestions or decisions to consumers, thereby satisfying consumers' self-identity and needs, generating a certain degree of psychological ownership, and satisfying their interaction motivation. Psychological ownership is equivalent to ownership at the perceptual level, reflecting the emotional attachment between users and AI agents, that is, the emotional support perceived by users during use. On the other hand, AI agents can provide emotional care to users by proactively capturing changes in the environment or changes in users' emotions. For example, the smart agent in a mobile phone can detect changes in the weather and proactively remind users to bring an umbrella before going out. These emotional supports and care make users feel that the AI ​​agent is caring and warm, and can arouse users' positive emotional perception. Therefore, the following hypothesis is proposed:

[0048] H2b. The autonomy of artificial intelligence agents positively affects consumers’ emotional perception.

[0049] Artificial intelligence agents can collect a large amount of data information by monitoring and controlling users' daily lives, and use this information to improve the ability to perform tasks independently (i.e., autonomy), which may cause users to perceive intrusion while improving service efficiency. In addition, a high degree of autonomy may make users feel out of control, which may lead to concerns about the restriction of rights, and users may think that such a system is at risk of intrusion. For example, during web browsing, web pages will automatically pop up ads based on consumers' browsing history, and users may think that pop-up ads are intrusive. Therefore, we propose the following hypothesis:

[0050] H2c. The autonomy of AI agents positively affects consumers’ perception of intrusiveness.

[0051] 2.2 Learnability and Consumer Perception

[0052] The learning ability of artificial intelligence agents means that the services or suggestions they provide are not stereotyped, but personalized services and practical suggestions given after adapting to environmental changes and interpreting user preferences. These services and suggestions reduce the information asymmetry of users themselves, reduce potential costs, and improve the efficiency and accuracy of problem solving. For example, AIA in the fields of commercial services and medicine can rely on deep learning to identify environmental changes with high precision, thereby improving the accuracy of task execution. Strong learning ability can enable users to clearly feel the reliability and usefulness of information technology, and may generate a positive perception of the functionality of artificial intelligence. Therefore, we propose the following hypothesis:

[0053] H3a. The learnability of artificial intelligence agents positively affects users’ functional perception.

[0054] The learning ability of AI agents is reflected in giving corresponding responses based on user preferences through their voice input or other forms of emotional expression, providing understanding and encouragement to users. For example, when users express their feelings to AIA, AIA can provide comfort or entertainment like a friend. This kind of emotional support obtained during the interaction can make users feel happy, comfortable and belonging, and positively affect their emotional perception. Therefore, we propose the following hypothesis:

[0055] H3b. The learning ability of artificial intelligence agents positively affects users’ emotional perception.

[0056] The learning ability of artificial intelligence agents is based on a huge database generated during the user's use process. It is necessary to continuously collect more data information from users and use this information to provide personalized services tailored for them. However, a high degree of personalization will make users aware that their personal information is being used, which makes them feel intrusive. For example, some advertisements recommended according to preferences and unconscious voice activation will make users feel dissatisfied and worried that their privacy has been violated. The improvement of learning requires more and more personal data to be collected and used, which may be regarded by users as a leakage of personal information, and they may suffer negative consequences in the future. These concerns will enhance users' perception of intrusion. Therefore, we propose the following hypothesis:

[0057] H3c. The learning ability of artificial intelligence agents positively affects users’ perception of intrusiveness.

[0058] 3. User perception and willingness to use AI agents; the social support perceived by users during their interactions with non-human entities will ultimately affect their subsequent behavioral intentions; functional perception is closely related to personal services, which is reflected in completing commands or tasks issued by users and solving problems faced by users. The technical advantages that users directly feel when using AIA help reduce information redundancy and can make users believe that intelligent agents are reliable and fair enough. Trust in non-human entities can directly affect people's willingness to continue interacting.

[0059] Emotional perception involves interaction, relationship, and emotional companionship with AI agents. The warmth perceived by users when using AIA can make them feel relaxed, comfortable, and happy, thereby maintaining and strengthening their trust and reliance on AIA. This trust and reliance may enhance users' sense of belonging and intimacy with AI agents, generating higher emotional value. Studies have shown that the psychological or emotional value enhancement effect significantly enhances users' continued use intention. Therefore, we propose the following hypothesis:

[0060] H4a. When interacting with an AI agent, users’ functional perception has a positive impact on their continued use intention.

[0061] H4b. When interacting with artificial intelligence agents, users’ emotional perception has a positive impact on their continued use intention.

[0062] Users' perception of the intrusiveness of AI agents is reflected in the fact that intelligent technology relies on collecting a large amount of user personal information. While users enjoy the convenient services brought by AI, they also see the threats brought by intelligent services. When users feel that their privacy is violated and they lose control of their personal information to a certain extent, they will feel worried or uncomfortable about intrusiveness. The negative emotions generated by intrusiveness will induce users to have negative interactive experiences, which may have a negative impact on users' subsequent behavior, that is, resisting or avoiding the use of AI agents.

[45] Therefore, we propose the following hypothesis:

[0063] H4c. When interacting with AI agents, users’ perceived intrusiveness has a negative impact on their continued use intention.

[0064] Figure 2The results of the structural model test were summarized. By explaining 49.6% and 53.4% ​​of the variance respectively, it was found that autonomy and learning had a positive impact on functional perception and emotional perception, thus supporting H2a, H2b, H3a, and H3b. By explaining 2.5% of the variance, it was found that autonomy had a positive impact on intrusive perception, while learning did not significantly affect intrusive perception, thus supporting H2c and rejecting H3c. The two dimensions of autonomy, learning, and psychological perception of artificial intelligence agents and the intrusive perception of equity theory jointly explained 50.6% of the variance in continued use intention. From the results, it can be seen that autonomy will negatively affect continued use intention, learning can directly positively affect use intention, functional perception and emotional perception have a positive impact on use intention, and intrusive perception has a negative impact on use intention, thus supporting H1b, H4a, H4b, and H4c, and rejecting H1a.

[0065] The impact of the autonomy and learning ability of artificial intelligence agents on users' continued willingness to use the technology was comprehensively examined from the three dimensions of functional perception, emotional perception, and intrusive perception.

[0066] First, the results show that autonomy and learning can enhance users' functional and emotional perception (H2a, H2b, H3a, H3b). In terms of functional perception, highly autonomous AIA can greatly help users complete their tasks without excessive user intervention and worry, thus improving work efficiency; highly learning AIA can provide users with highly personalized experiences and constantly updated information bases. In terms of emotional perception, AI agents with high autonomy and high learning can show strong empathy and communication skills, provide companionship to users, make users feel warm, and receive strong emotional value.

[0067] The results also show that autonomy will enhance users' perception of intrusion (H2c), while on the contrary, learning has no significant effect on users' perception of intrusion (rejecting H3c). AIA with strong autonomy will collect users' personal information and data in the process of daily task execution, which obviously will also have certain privacy and security risks. At the same time, it may also make users over-rely on intelligent decision-making and lose control of their own choices, which will make users feel uneasy about their lives being disturbed. Deep learning technology can enable intelligent agents to solve many problems that are difficult for humans to solve. Equity theory aims to explain users' views on perceived benefits and consumption during the interaction process. AIA can reduce their concerns about privacy by providing users with more benefits. The learning of artificial intelligence agents is more reflected in the enhancement of system performance, and does not involve the collection and use of user information. Users do not need to worry about their privacy security being threatened.

[0068] Second, users' functional and emotional perceptions of AIAs will enhance their continued use intention (H4a, H4b), while intrusive perceptions will reduce their continued use intention (H4c). The functional advantages and emotional value of highly autonomous and highly learnable AIAs will bring users higher trust and emotional attachment. Trust and intimacy can enhance users' continued use intention. However, users will also suffer from information intrusions brought by highly intelligent AIAs, including privacy leaks, information cocoons, algorithmic bias, and decision-making errors, which will make users feel uneasy, and then resist AIAs and reduce their continued use intention.

[0069] Third, inconsistent with the expected hypothesis, the autonomy of AI agents will directly reduce users' continued use intention (H1a). In the absence of other factors, high autonomy may cause users' concerns or distrust. Users may worry about losing control of AI agents, or worry that agents with too high autonomy may make decisions that do not meet their expectations or interests. Under the influence of intrusive perception, users may feel the lack of autonomous control and the risk of information being invaded. However, the positive feedback brought by functional perception and emotional perception alleviates users' uneasiness, thereby enhancing users' continued use intention. On the other hand, the learning ability of AI agents can directly enhance users' continued use intention (H1b). The powerful learning ability of AI agents can provide users with more accurate services or suggestions by learning user preferences and adapting to environmental changes, thereby enhancing user trust. The improvement of learning level will not cause users to perceive intrusion, but will only enhance functional perception and emotional perception.

[0070] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for analyzing user's continued use intention by an artificial intelligence agent, characterized in that: The user's perception of AI agents is divided into functional perception, emotional perception and intrusive perception; the user's continued use intention is analyzed from the following perspectives: a. The intelligence of artificial intelligence agents and users' continued willingness to use them; for users to accept and continue to use such products; b. Perceptual theory of mind and fairness theory; users assess the value of AI agents based on their social interactivity and personality traits; When users interact with an AI agent, they tend to perceive it as human-like behavior; c User perception and intention to use AI agents; the social support perceived by users during their interactions with non-human entities will ultimately affect their subsequent behavioral intentions; functional perception is closely related to personal services, which is reflected in completing commands or tasks issued by users and solving problems faced by users.

2. The method for analyzing user's continued use intention by an artificial intelligence agent according to claim 1, characterized in that: aIncludes autonomy and willingness to continue using and learning and willingness to use; specifically as follows: H1a. The autonomy of AI agents positively affects users’ continued use intention; H1b. The learning ability of artificial intelligence agents positively affects users’ continued use intention.

3. The method for analyzing user's continued use intention by an artificial intelligence agent according to claim 1, characterized in that: b includes autonomy and user perception and learning and consumer perception; the details are as follows: H2a. The autonomy of AI agents positively affects users’ functional perception; H2b. The autonomy of artificial intelligence agents positively affects consumers’ emotional perception; H2c. The autonomy of AI agents positively affects consumers’ perception of intrusiveness; H3a. The learning ability of AI agents positively affects users’ functional perception; H3b. The learning ability of artificial intelligence agents positively affects users’ emotional perception; H3c. The learning ability of artificial intelligence agents positively affects users’ perception of intrusiveness.

4. The method for analyzing user's continued use intention by an artificial intelligence agent according to claim 1, characterized in that: c includes psychological or emotional value enhancement effects that significantly enhance users' continued use intention; specifically, as follows: H4a. When interacting with an AI agent, users’ functional perception has a positive impact on their continued use intention; H4b. When interacting with an AI agent, users’ emotional perception has a positive impact on their continued use intention; H4c. When interacting with AI agents, users’ perceived intrusiveness has a negative impact on their continued use intention.

Images