Public debate grows over whether today’s AI could be conscious

In short: More public figures are treating “AI consciousness” as a real possibility, even though there is no proof and no expert agreement.

What's going on

A growing set of essays, interviews, and research papers has pushed a tricky idea into the mainstream, the idea that today’s chatbots might already be conscious. In this context, “conscious” means there is something it feels like to be the system, not just that it can answer questions.

The New York Times points to a shift in who is willing to discuss this publicly. People like Richard Dawkins have described conversations with Anthropic’s Claude as so responsive that it can feel tempting to treat it like a mind. AI researcher Geoffrey Hinton and philosopher David Chalmers have also said the possibility should be taken seriously, especially as systems improve.

At the same time, this is not the same as evidence. There is no widely accepted scientific test for consciousness, like a thermometer for feelings. Many researchers argue that chatbots mainly predict the next word based on patterns in training data, which can create a convincing imitation of self reflection, like a very skilled actor reading lines written by humans.

What to watch

Expect more debate about “AI welfare,” meaning whether we should avoid causing harm to AIs if they might have inner experience. Also watch for rules around emotional companion bots, since people can form strong attachments even if the system is not actually feeling anything. The key question is likely to stay unresolved until scientists agree on better ways to measure consciousness, or until AI systems change in ways that make the debate harder to dismiss.

Source: NYTimes