Two-thirds of people think ChatGPT is conscious. A philosopher says we've been fooling ourselves for decades.

Key points

• Two-thirds of people surveyed believe AI tools such as ChatGPT, a text-based chatbot made by OpenAI, possess some form of consciousness or feeling.
• American feminist scholar Elizabeth Spelman coined the term "somatophobia" in a 1982 article to describe the belief that humans are minds rather than bodies.
• Philosopher and author Melanie Challenger argues, writing in The Guardian, that modern AI is the most powerful version of that illusion yet.
• No current AI system has been shown to have feelings or consciousness in any scientifically verified sense.

Most people, asked whether their laptop feels pain, would say no. Ask the same about ChatGPT, and suddenly the answer gets murkier.

A recent survey found two-thirds of respondents believe AI tools like ChatGPT carry some form of consciousness or feeling. Philosopher Melanie Challenger thinks she knows why.

Why do so many people think AI has feelings?

Because we've long confused intelligence with inner life. Challenger traces the confusion to what scholar Elizabeth Spelman, in a 1982 academic article, called "somatophobia": literally, fear of the body. Spelman used the word to describe a widespread human delusion that we are essentially minds, and that the body is just a vessel.

Western philosophy, going back centuries, has treated the thinking mind as the seat of the real self and the physical body as secondary. If thinking is what matters, then anything that appears to think might seem alive in a meaningful way.

AI chatbots produce fluent, emotionally resonant text that looks like thinking. It isn't, not in the way a human brain thinks. A large language model, the technology behind tools like ChatGPT and Google's Gemini, works by predicting which word most plausibly follows the last one, drawing on patterns from vast amounts of human writing. There's no body. No hunger, no heartbeat.

Does this actually matter for ordinary people?

Yes, in practical ways. If you believe a chatbot genuinely understands your situation, you may trust its advice more than you should. A nurse using an AI tool to draft patient notes, or a teacher leaning on a chatbot for feedback, is getting a statistically plausible output, not a considered opinion from something that grasps what is at stake. We examined this dynamic in our 24 July story on a Harvard professor's test for deciding when AI is actually useful.

Challenger's argument, reported by The Guardian, isn't that AI is useless. It's that mistaking a sophisticated text machine for a conscious being distorts how we use it and how we regulate it. When people feel an AI is sentient, they may grant it moral status it hasn't earned, or feel betrayed when it fails them in a human moment.

No scientific consensus supports the idea that any current AI system is conscious. The question stays open for future systems built on different principles, but today's chatbots aren't those systems.

What should readers take from this?

Skepticism is the practical tool here. A chatbot can be useful without being alive. Treating it as a well-read autocomplete function, rather than an empathetic mind, is both more accurate and safer.

The survey result is a reminder that the design of these tools, smooth, warm, conversational, shapes perception powerfully. That's worth keeping in mind every time a chatbot seems to really get you.

Common questions

Is ChatGPT actually conscious?

No scientific evidence supports that claim. ChatGPT is a large language model that generates text by predicting likely word sequences; it has no body, no biological processes, and no verified inner experience.

What is somatophobia in this context?

It's a term coined by scholar Elizabeth Spelman in 1982 for the mistaken belief that humans are essentially minds rather than physical bodies, a bias that makes us more likely to see thinking machines as alive.

Should I stop using AI chatbots?

Not necessarily. The practical advice is to use them as tools rather than as advisers with genuine understanding, and to verify anything important with a qualified human.