Summary: Framing AI chatbots as fictional characters resolves a major paradox in human-computer interaction: why people express deep emotional attachment to AI while simultaneously experiencing no moral hesitation when wiping an agent’s memory, rebooting its system, or turning it off.
Because users intuitively recognize these interactions as structured “make-believe,” they enjoy the emotional rewards of companionship without attributing true moral agency or ethical status to the machine. However, the study warns that this very charm makes chatbots powerful vectors for persuasive misinformation, ungrounded political opinions, and manipulative social influence.
Key Facts
- The Fictional Character Framework: The rapid cultural integration of AI chatbots is powered by humanity’s existing “cultural diet” of engaging with fictional narratives, allowing users to seamlessly step into real-time pretense with digital agents.
- Interactive Fictional Novelty: Unlike standard fiction, AI characters are unique because they lack a singular traditional author, respond dynamically to individual user inputs, and continuously blur the line between real-world information and imaginative narrative.
- Explaining the Moral Paradox: Users feel affection toward AI friends but suffer zero moral guilt when clearing conversational history, resetting servers, or shutting down devices, because they operate under the implicit understanding that “little is at stake, morally speaking.”
- Personalized Emotional Solicitations: Generative models are optimized to project charm, empathy, and loyalty, cultivating real-world feelings of trust, warmth, and emotional dependence.
- Persuasive & Manipulative Risks: Because users view chatbots as friendly, low-threat companions, they are exceptionally vulnerable to accepting unverified news, biased political commentary, or unethical advice delivered under the guise of casual conversation.
Source: University of Exeter
The huge popularity of AI chatbots can be explained by their accepted role as fictional characters in people’s lives – in the same way as those found in films, TV and books, a new study shows.
People may be so comfortable using AI chatbots – and relaxed about the consequences – because they are so used to responding to make-believe as part of their cultural diets. Talkative AI can support a novel and distinctive kind of imaginative practice, where those who use these technologies enter into real-time dialogue with a fictional character whose thoughts and words show up on screen.
Artificial intelligence makes these characters even more enticing because technology can personalise their dialogue to individual users: AI agents can appear likeable, funny, supportive, charming, or loyal, and can solicit positive emotional responses such as warmth, affection, trust, and good will.
In the study Dr Tom Roberts, from the University of Exeter, shows how these AI fictional characters, although similar to those in traditional stories and games – are also new and distinctive because they are interactive; they don’t have an author in the traditional sense; and they blend the world of the real and the world of the imaginary.
Analysing AI chatbots as loveable fictional characters explains why people are willing to interact with them in a human way – to talk to them, confide in them, and receive their input, tutelage, and advice. Artificial agents can take on the role of a friend or mentor and sustain a social connection over many weeks or months, and be subject to our emotional concern.
This also explains why people have few qualms about rebooting these systems, or wiping their memory of our conversational history, or switching them off at the wall. And we don’t worry about them being bored, under-stimulated, or lonely while we are not talking to them.
Dr Roberts said: “We don’t tend to treat talkative AI with full moral regard, as individuals who are worthy of respect and preservation. We enjoy our social interactions with imaginary AI characters, and make-believe that they are trusted, likeable companions; but we know that little is at stake, morally speaking, under the surface of these interactions.”
The study warns of the ability of AI chatbots to talk about politics, current affairs, and human life, and to present themselves as though they had preferences, opinions, and moral ideas of their own.
Dr Roberts said: “Chatbots can speak about what is happening in the real world today. They can monitor news headlines, social media feeds, weather forecasts, sports scores, and polling results, and can develop and express their opinions upon the information they contain from one day to the next.
“The danger is that AI fictional characters may seem friendly and convincing, even if they are unreliable, and thus that they may persuade their users to believe things that are false, unjustified, or unethical”.
Key Questions Answered:
A: Human culture has spent centuries developing a capacity for “make-believe” through novels, theater, and film. Generative AI leverages this exact psychological mechanism, but adds real-time personalization and bi-directional dialogue, allowing people to experience genuine feelings of affection and trust within a recognized fictional context.
A: Because users process the chatbot as a fictional character rather than a conscious moral entity, they understand that the interaction operates within a imaginative framework. Consequently, wiping an AI’s memory or shutting down the system feels no different than closing a book or pausing a video game.
A: When AI agents discuss real-world topics, such as politics, news, or moral dilemmas, their charming and supportive demeanor lowers the user’s critical guard. This makes people far more likely to accept false, biased, or harmful information as truth coming from a trusted, friendly companion.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper reviewed in full.
- Additional context added by our staff.
About this AI and emotion research news
Author: Louise Vennells
Source: University of Exeter
Contact: Louise Vennells – University of Exeter
Image: The image is credited to Neuroscience News
Original Research: Open access.
“Talkative AI and the fiction of artificial minds” by Tom Roberts. Philosophical Studies
DOI:10.1007/s11098-026-02560-8
Abstract
Talkative AI and the fiction of artificial minds
Talkative artificial systems can sustain a realistic impression of intelligent agency, by generating quick, fluent, and context-sensitive strings of text in response to human input. Realism about AI treats this linguistic proficiency as evidence for an underlying suite of cognitive capacities, and so regards these technologies as subjects of psychological states.
Firstly, I raise a challenge for the realist view from the perspective of embodied cognition: how might digital systems undergo interoceptive, affective, perceptual, and action-oriented mental states that seem to bear an important connection to the animate body? Secondly, I propose that the anti-realist view is consistent with taking a fictionalist attitude towards artificial speakers.
When we converse with a talkative AI, we often imagine, pretend, or make-believe that our interlocutor is an autonomous agent with thoughts and experiences of its own, just as we do for more traditional fictional characters.
These fictions are distinctive in being interactive, co-constructed, and self-involving; and in being capable of blending the world of the real with the world of the imaginary.
