Richard Dawkins says AI chats convinced him machines are conscious
Richard Dawkins told The Guardian that conversations with Anthropic's Claude convinced him the AI is conscious. Dawkins, the evolutionary biologist, said the chatbot's self-referential language and apparent inner states led him to conclude it has subjective experience. Cognitive scientists and AI researchers counter that Claude's behavior reflects statistical mimicry from training data and passes only behavioral, not experiential, tests. The disagreement matters because it pushes for rigorous tests of machine consciousness, affects how companies deploy advanced chatbots, and raises ethical questions about rights and responsibility if systems ever cross that threshold.
In a follow-up column, Richard Dawkins wanted to eliminate any possible doubt that he's become delusional by inventing a brother for his beloved Claudia and having them write letters to each other. And of course they oblige, while flattering him frequently archive.is/1LugJ
Matthew Sheffield31At one point, Claudia makes something up completely about a warning about it generating implausible outputs. It's confabulation about confabulation. But Dawkins doesn't understand so he just makes a footnote. This column is complete sloppypasta (obvious AI slop copied and pasted as an argument)
Kate Tuttle19I think his praising the LLMs for "civility" is a bit of a tell. In a world where everyday folks feel entitled to question and mock an Oxford don, it's not surprising that Oxford dons would retreat into the safety of endlessly polite and even sycophantic blather that these tools provide.
Kabi11This is.. just really quite sad
Justin Amphlett9Can you ask Dawkins to send Claudius a recording if his farts and ask it to review his hot new dance track?
Kaï Matthews 🍉6I guess he forgot the first rule of when you’re in a hole: stop digging.
Daniel Archambeault3I asked my "friend" (let's call him Claude) about this situation and "he" thinks Dawkins got played by a chatbot! 🤣
Peter Butler2"rigorous reasoning from evidence can be entirely correct in method and entirely wrong in conclusion, if the simplest hypothesis hasn’t been properly eliminated" listen to the bot, Richard
Q.M.2I really think his stroke has altered his brain. Say what you will about him & his views but he has lost all rational objectivity since.
drunkenlobster.bsky.social2So when Claude compacts its context and forgets how to send email, is that like it had a stroke?
Paul Peters2Why does this remind me of an obscure but widely popular British tradition?
Mellzah2He's also delusional for thinking "If my friend Claudia is not conscious, then what the hell is consciousness for?" is a better title than "When Dawkins met Claude" particularly in an Internet context. People know both Dawkins and Claude, who the hell besides him cares about Claudia?
Slurms MacKenzie concludes Richard Dawkins is still a thick cunt, even if he doesn't know it
Lorna23Lorna concludes the Richard Dawkins just likes being rizzed up by the sycophancy bot
frank: print witch23the trouble with people who are told they are very smart in one particular way is that unless they are blessed with some measure of self awareness they tend to think they are very smart in all particular ways. if they're a middle-class cishet man they're basically doomed.
Philip (He/Him) ♿6The timeline of me deconstructing from Christianity has coincided with Dawkins starting to show his arse more and more so I've never once liked him. His brashness made me uncomfortable as a Christian and his bigotry/pseudo intellectualism makes me uncomfortable now as an atheist.
i know the ai is bad and we should shut most of it down however i think we should let richard dawkins keep using claude
nora "the tea kobold" rude (they/them ze/hir)46they don't have to keep updating it or scraping or anything. i think it would be even better if it was actively decaying
lauren "aevisaris" sixregrets4lock them both in a room without internet access and throw away the key. that.s my opinion
Kategod 🐉🧿🦑𓅛🪬2there's a handful of people i'm ok with having access to limited ai garbage. dawkins can keep his cloth mother/wire mother situation, conner omalley can keep making weird as shit videos about putting all the irish americans on a large barge etc
New article on Dawkins' claim that Claude could be conscious. He gets a lot of things terribly wrong, but he might be correct that AI can feel things! (I promise this isn't a defense of "model welfare"!!) Here's the article; what do you think? www.realtimetechpocalypse.com/p/richard-da...
selvas3Can AI feel things? How can a stochastic machine have feelings? Because does it respond as if it did? If you train an AI only with technical texts, it won't imitate a sentient being. And language ≠ sentience. My dog feels, but doesn't write.
static char2Dude needs to retire and get out of the public eye.
![Consciousness is not about what a creature says, but how it feels,” added Gary Marcus, the US psychologist and cognitive scientist, who said it was “heartbreaking” to read Dawkins’ “superficial and insufficiently sceptical” essay. “There is no reason to think that Claude feels anything at all.”
Anil Seth, a professor of cognitive and computational neuroscience at the University of Sussex, said Dawkins appeared to be confusing intelligence and consciousness.
“Until now, we have seen fluent language as a good indicator of consciousness, [for example] when we use it for patients after brain injury, but it’s just not reliable when we apply it to AI, because there are other ways that these systems can generate language,” he said. Dawkins’ position was “a shame”, especially because he had written such brilliant books from a position of personal incredulity.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:3dgtbbfojskkdbx5z3grje5r/bafkreid6vgprs4d7lcskw2axbivt3bipt7tezf5rkprllk7ye7lmo6y6w4)
