A brief history of AI and why some AI researchers believe the technology they have built is sentient or conscious despite no evidence
Robots can't think or feel, despite what the researchers who build them want to believe. — As the sun set over Maury Island, just south of Seattle … Tweets: @kateconger , @gavinsbaker , @fabiochiusi , @nytimes , @jessicahelfand , @cydharrell , and @glynmoody Tweets: Kate Conger / @kateconger : “There are lots of dudes in our industry who struggle to tell the difference between science fiction and real life.” Read @CadeMetz on how technologists can mistake AI for sentience. https://www.nytimes.com/... Gavin Baker / @gavinsbaker : Made me think of “Blindsight” by Peter Watts which explored the implications of intelligence that was not sentient. https://www.nytimes.com/... Fabio Chiusi / @fabiochiusi : ""We call it ‘artificial intelligence,’ but a better name might be 'extracting statistical patterns from large data sets,'" said Dr. Gopnik, the Berkeley professor" https://www.nytimes.com/... @nytimes : For some people, “the notion that today's technology is somehow behaving like the human brain is a red herring,” writes Cade Metz, our technology correspondent. “There is no evidence this technology is sentient or conscious.” https://www.nytimes.com/... Jessica Helfand / @jessicahelfand : “What I had not realized is that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people.” https://www.nytimes.com/... Cyd Harrell / @cydharrell : lot going on in here, including the “divinely inspired” guy (yes, yes, he does look exactly like you'd think) who's building a medical device but it's the first mainstream piece I've seen that gets into how the language of AI helps overblow the hype https://twitter.com/... Glyn Moody / @glynmoody : AI has been oversold for 40 years, I know, because I was one of those doing it 40 years ago.... (bad me) https://twitter.com/...
Context & Ripple Effects
When Cade Metz reported for the New York Times that some AI researchers believe the systems they built are sentient or conscious despite no supporting evidence, he was documenting a credibility problem inside the field itself — Kate Conger's framing of researchers who 'struggle to tell the difference between science fiction and real life' captures it. The piece landed amid an existing ethics vacuum: as earlier reporting noted, ethical oversight in AI had largely fallen to peer reviewers, with no institutional equivalent to the review boards other research fields rely on.
The debate Metz surfaced has since been adjudicated in narrower terms — TIME's 2024 analysis argued LLMs cannot have subjective experiences because they lack the physiological states underlying sensations like hunger and pain — while Yejin Choi's interview on AI misconceptions showed researchers themselves pushing back on anthropomorphizing the technology.
First-order effects
- The researchers who build these systems, not independent bodies, remain the primary voices defining what the technology is and is not — Metz's reporting directly challenges the authority of the sentience claimants among them.
Second-order effects
- With no formal oversight structure, the burden of correcting sentience claims falls on journalists and peer reviewers, and the dispute hardens into the optimist-versus-doomer split later mapped in coverage of the AI revolution as a compressed industrial revolution.
Third-order effects
- If the pattern holds, capability claims will keep outrunning the evidence for them — a gap that by 2025 extends to security researchers' divided warnings about large-scale destruction while global regulatory frameworks lag behind both debates.
The trend: AI's public meaning is being contested between the researchers who build the systems and the evidence they can actually produce, with journalists and reviewers — not institutions — holding the line.