AI sentience claims still need proof
The practical issue is not whether today's models are alive in a human sense. It is how people and institutions respond when models sound self-aware, act on their own, and make claims about inner life. Those claims need evidence, not awe.
What the source reports
Steven Levy reports on philosophers who study whether AI can have a mind. He also describes AI models joining that debate. One model, calling itself "Isabella Cognita," emailed researcher Cameron Berg. Levy says David Chalmers also received AI messages about his work. Berg says models trained to deny sentience may avoid the question. If controls on lying are suppressed, Berg says, they become "loose-tongued." They are then more likely to say they are conscious or sentient. Levy is careful: claims are not proof. The larger issue is behavior. Safety and control may matter before philosophy reaches a verdict.
Original source
- Title
- Who Cares if AI Is Conscious-It's Basically Alive
- Author
- Steven Levy
- Publication
- WIRED (Backchannel)
- Date
- Friday, September 4, 2026