Summary: Casting AI chatbots as fictional characters solves a major paradox in human computer interaction: why people express deep emotional attachment to AI while at the same time having no moral qualms when erasing an agent's memory, restarting its program, or shutting it down.
Because users see these interactions as programmed “pretend,” they enjoy the emotional rewards of friendship without attributing any real moral agency or moral status to the machine. However, this study warns that this very charm makes chatbots powerful vectors of persuasive misinformation, baseless political opinions, and perverse social influence.
Important Facts
- Fictional Character Framework: Rapid cultural integration of AI chatbots is powered by people's existing “cultural appetite” for engaging with fictional stories, allowing users to seamlessly engage in real-time simulations with digital agents.
- Interactive Fictional Novelty: Unlike conventional fiction, AI characters are unique in that they lack a single traditional author, respond dynamically to each user's input, and constantly blur the line between real-world information and imagined narrative.
- Explaining the Moral Paradox: Users feel love for AI friends but not guilt when they delete chat history, rearrange servers, or close apps, because they work under the clear understanding that “less is at stake, morally speaking.”
- Personalized Emotion Applications: Productive models are developed to project attractiveness, empathy, and credibility, cultivating real-world feelings of trust, warmth, and emotional dependence.
- Persuasive and Insidious Risks: Because users view chatbots as friendly, low-risk companions, they are more vulnerable to accepting unverified news, biased political commentary, or unethical advice delivered under the guise of casual chat.
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 that they are found in films, TV and books, new research shows.
People may be more comfortable using AI chatbots — and comfortable about the results — because they are more accustomed to responding to make-believe as part of their cultural diet. Talkative AI can support a novel and different kind of cognitive practice, where those who use this technology enter into a real-time conversation with a fictional character whose thoughts and words appear on the screen.
Artificial intelligence makes these characters even more attractive because the technology can personalize their conversation to individual users: AI agents can appear likeable, funny, supportive, likeable, or trustworthy, and can invoke positive emotional responses such as warmth, love, trust, and goodwill.
In a study Dr Tom Roberts, from the University of Exeter, shows that these fictional AI characters, although similar to those in traditional stories and games – are also new and different because they interact; they have no author in the traditional sense; and they mix the real world with the imaginary world.
Analyzing AI chatbots as lovable fictional characters explains why people are willing to interact with them in a human way – talking to them, confiding in them, and getting their ideas, instruction, and advice. Artificial agents can take on the role of a friend or mentor and maintain contact with people over many weeks or months, and be subject to our emotional concerns.
This also explains why people have few doubts about restarting these systems, or clearing their memory of our chat history, or turning them off the wall. And we don't worry about them being bored, overstimulated, or lonely when we don't talk to them.
Dr Roberts said: “We don't tend to treat talking AI with full morals, as human beings worthy of respect and preservation. We enjoy our interactions with imaginary AI characters, and we make them believe that they are loyal, dear friends; but we know that there is little at stake, morally speaking, under these interactions.”
This study warns of the ability of AI chatbots to talk about politics, current affairs, and people's lives, and present themselves as if they have their own likes, opinions, and moral views.
Dr Roberts said: “Chatbots can talk about what's happening in the real world today.
“The danger is that fictional AI characters may appear friendly and convincing, even if they are not trustworthy, and thus may cause their users to believe false, unjust, or immoral things”.
Important Questions Answered:
A: Human culture has spent centuries developing the power of “make believe” through novels, plays, and film. Generative AI uses this precise psychological approach, but adds real-time personalization and two-way dialogue, allowing people to experience real feelings of love and trust within a familiar fictional context.
A: Because users consider the chatbot as a fictional character rather than a moral entity, they understand that the interaction operates within an imagined framework. So, wiping the AI's memory or shutting down the system doesn't feel like closing a book or pausing a video game.
A: When AI agents discuss real-world topics, such as politics, news, or ethical dilemmas, their positive and supportive behavior lowers the user's vigilance. This makes people more likely to accept false, biased, or damaging information as truth from a trusted, friendly friend.
Editor's Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper is fully revised.
- More content is added by our staff.
About this AI and emotion research issues
Author: Louise Vennells
Source: University of Exeter
Contact person: Louise Vennells – University of Exeter
Image: Image posted in Neuroscience News
Actual research: Open access.
“Talkative AI and the myth of artificial intelligence” by Tom Roberts. Philosophical Studies
DOI:10.1007/s11098-026-02560-8
Abstract
Talkative AI and the myth of artificial intelligence
Artificial speaking systems can maintain the original vision of an intelligent agency, by generating fast, fluent, and context-sensitive text strings in response to human input. Realism about AI takes these language skills as evidence of a basic set of cognitive abilities, and considers these skills to be subjects of mental states.
First, I raise a challenge from the perspective of realism from the perspective of embodied psychology: how is it possible for digital systems to deal with subjective, affective, perceptual, and action-oriented mental states that seem to have an essential connection with the living body? Second, I suggest that the anti-realist view is accompanied by the assumption of a fictional status by artificial speakers.
When we interact with a talking AI, we tend to imagine, pretend, or believe that our speaker is an independent agent with real thoughts and experiences, as we do with many traditional fictional characters.
These myths are unique in that they are interactive, co-constructed, and inclusive; and being able to bridge the real world with the symbolic world.
