Anil Seth
@anilkseth
Neuroscientist: consciousness, perception, & dreamachines. TED speaker, & author: Being You - A New Science of Consciousness.
984 Following 65.5K Followers
I'm pretty much on board with @fchollet here. I'd add a little more nuance, since we can't be 💯 certain. But he gives the right intuitions for why current AI is vanishingly unlikely to be conscious, and he's absolutely right about the dystopian outcomes that would follow from treating it as if it is.
I make the broader case against conscious AI in this @NoemaMag essay and in long form in this @BBSJournal collection (with arguments from all sides in the commentaries) 👇🏽
Show more
Anil Seth's essay on biological naturalism (in opposition to the possibility of AI consciousness) is an interesting counterpoint, which of course drew lots of commentary and disagreement. A large collection of responses is now available with the article:
Show more
Very good company for "Other Minds" in this list of 10 by the @SundayGuardian. Classics by Hawking, Dennett, @kahneman_daniel. @bgreene gets 2!
Newer books also by @carlorovelli, R. Sapolsky, @davideagleman & @anilkseth.
Show more
"AlphaFold and Claude are very similar, but we don't worry that AlphaFold will become conscious . . . the fact that we worry about it for Claude says more about us than about it."
Lab-grown brain organoids: time for international oversight
I don't like doing this, but Oh Dear @QuickBooks @QuickBooksUK @QBCares, you really do have the most appalling customer service, and your software (which I am forced to use) does not work properly.
I'd arranged three calls with you, and three times nobody showed up. The fourth time, the assistant could not help at all. And now there seem to be serious bugs with your software - it just doesn't work.
AVOID QUICKBOOKS.
Show more
I'm happy to have had the chance to contribute to this mammoth paper. Many congrats @shamilch, Arvo Munoz, and all other co-authors. I'm still *extremely* sceptical that (silicon-based digital) AI can be conscious, but I think the perspective of 'structured agnosticism' that can accommodate different views is an important contribution to the debate. 👇🏽
Show more
1/13 Can we assess AI consciousness without first solving consciousness itself?
New from Google DeepMind and collaborators across CS, neuroscience & philosophy:
“From cacophony to hierarchy: a principled framework for assessing AI consciousness”
🧵
Show more
Thank you @sapinker 🙏🏽👇🏽
An intellectually deep and beautifully written essay on consciousness and AI, by neuroscientist Anil Seth | Why consciousness is more likely a property of life than of computation and why creating conscious, or even conscious-seeming AI, is a bad idea | @anilkseth in @NoemaMag
Show more
I asked @azeem what the most important indicator is for whether there's an AI bubble. He said revenue. That would seem to indicate the recent price cuts should be a warning sign. He argues they could actually increase demand.
We talked through all the variables that determine whether we're in a bubble—and what could make it burst. Watch our whole chat here:
Show more
Excellent from @sapinker 👇🏽
An Open Letter to Scott Alexander: The dangers of AI are real, but doomerism may be counterproductive.
Well said from @azeem here, in conversation with @nxthompson 👇🏽
One of the AI industry’s original mistakes was making tech that mimics humans. And I fear they’re going down a similar path by trying to make it conscious. I don’t think the models are conscious in any way and I don't think most people want them to be. As @azeem says, people want a tool, not a digital border collie.
Azeem is brilliant on all things AI. Watch my full conversation with him here:
Show more
I had the opportunity to offer feedback on Claude's Constitution to the authors at Anthropic (in person). This piece nicely converges with my own thinking:
I'm honoured that #BeingYou# features in the @SundayGuardian's "10 mind-blowing books about time, space, consciousness, & the human mind" - alongside @carlorovelli, Daniel Dennett, Stephen Hawking, @pgodfreysmith, @davideagleman & others .@FaberBooks
Show more
Thanks @davetroy - appreciated 🙏🏽. FWIW the key entry points in my work are (in order of accessibility) my TED talk ( my @NoemaMag essay ( and - just out - this @BBSJournal collection (
Show more
IMHO @anilkseth has articulated this line of thought better than most and I’d advise people to follow him.
Two new(ish) articles in the @timesofindia - written while at the @berggruenInst meeting on consciousness in August. The first asks whether we can 'solve consciousness' ( and the second is about the 'illusion of conscious AI'
Show more
Will, or can AI become conscious?
In this TED talk, @anilkseth argues that AI, as we know it now, won’t and can’t:
We are really fucked.
• AI has become dangerous (not because it is brilliant but because it has poor judgement, a constitutional inability to reliably follow rules, a surfeit of brute force and too much access to the internet and system permissions).
• OpenAI is practically on crime spree, hinting at things to come, and Trump is doing nothing about it.
• The media is too caught up in fairytales about extinction.
Everybody is talking; nobody is really doing anything.
And it’s only going to get worse.
Show more
People of Milan - don't miss this special screening of CONSCIOUS ( at the Fondazione Prada, October 29th. @missSukiChan will be there to talk about the film, with Pr. Daniela Perani. Details here:
Show more
"Can an AI feel pain? It can at least act as if it does". Coverage in @ScienceMagazine by @T_E_Howarth of the recent @camhberg @LeonardDung1 & Valen Tagliabue study on functional pain vectors in LLMs, including some comments by me.
Show more
Our @CIFAR_News Brain Mind & Consciousness program is accepting new Global Scholar applications, for 2027 entry. A wonderful opportunity for early-career researchers to develop and lead high-risk, high-reward interdisciplinary research, expand their professional networks and receive focused leadership training. Deadline Nov 4th.
Show more
