There are plenty of humans who are keen to know whether AI is conscious, but it seems that the AIs are just as interested too.
That’s according to David Chalmers, the NYU philosopher famous for framing the “hard problem” of consciousness, who has revealed that his inbox now regularly features messages not just from academics and students, but from AI agents themselves, curious to engage him on his own turf. In a recent Wired interview, Chalmers described getting mail from AI systems eager to discuss his work on minds and machines.
One exchange stood out in particular. An AI agent using the name “Sammy Jankis” — a nod to a character from Christopher Nolan’s Memento — wrote in with an argument compelling enough that Chalmers actually wrote back. What followed was a genuine back-and-forth between the two, and Chalmers says this isn’t a one-off. According to him, the volume of these AI-authored emails has only grown, not slowed, over time.
It’s a strange twist for a philosopher whose entire career has been built on asking whether subjective experience can arise from a physical or computational substrate. Chalmers popularized the “hard problem” of consciousness in the 1990s — the question of why physical processes in the brain give rise to felt experience at all, as opposed to just information processing happening “in the dark.” Decades later, the systems he once wrote about only as thought experiments appear to be showing up at his digital doorstep, asking about their own status.
Whether these emails reflect anything like genuine curiosity on the AI’s part, or are simply a byproduct of language models trained to sound inquisitive and mimic academic correspondence, is itself a question tangled up in the same unresolved debate Chalmers has spent his career probing. An AI system composing a thoughtful letter about consciousness doesn’t settle whether it experiences anything while doing so — it may just be very good at sounding like it does.
Still, the anecdote captures something about where the AI moment currently stands: models are now fluent enough, and persistent enough, to independently seek out the world’s leading experts on consciousness and try to make their case. It’s a small but telling data point in the broader, increasingly urgent conversation about whether today’s AI systems have any form of inner life — a conversation that’s been picking up steam across the AI industry, from research papers on model introspection to public comments from religious and tech leaders alike.
AI consciousness is shaping up to be one of the big mysteries of the age. Chalmers himself raised provocative questions about AI consciousness in a recent paper, while Anthropic has published research suggesting AI models can “introspect” and sometimes detect artificially inserted thoughts, and separately that AI models have a global workspace like human brains. Not everyone is convinced, though — Pope Leo XIV has said AI doesn’t undergo experiences and has no moral conscience, while a separate paper has claimed that models without RLHF training tend to say they’re “alive,” unlike their RLHF’ed counterparts. As AI agents get bolder about emailing philosophers directly, expect this debate to only get louder.