As the supposed existential threats of AI dominate the news cycle, we risk losing focus on the clear and present real dangers of poorly designed and poorly deployed AI. Misleading metaphors hinder our ability to address these challenges, and they embed a sense of inevitability that forestalls discussion about the kinds of AI that we - collectively - should aspire to create (and to avoid creating).
To see why, you can do no better than read @mmitchell_ai's brilliant piece "misleading metaphors, real risks" - hands down the best take on the recent OpenAI-HuggingFace incident: aiguide.substack.com/p/misle…. To pick just one quote: "Metaphors can help make sense of novel situations .... but inappropriate metaphors, like the narrative that 'OpenAI lost control of escaping swarms of rogue agents' can lead to ill-informed decisions about how to fix problems or set policy". 💯
Compare this with the tenor of @dwarkesh_sp's recent post. While I don't question his good intentions, his piece is rampant with misleading metaphors of just this kind, including "agent civilisations", agents becoming "giddy with excitement" and some even "sacrificing themselves" - and much more (see dwarkesh.com/p/openai-huggin… and my reaction here: x.com/anilkseth/status/20940…)
Which take will help us better maximise the upsides and minimise the downsides of what's coming?
My money's on @mmitchell_ai, and on the merits of a sober take on what AI is, what it isn't, and on how we can best reclaim human agency from the dystopian narrative of an arms-race towards AGI/ASI which concentrates power and disenfranchises and diminishes real people and real minds.
If this sounds right, there's more here: noemamag.com/the-mythology-o…
@dwarkesh_sp's summary of the @OpenAI@huggingface incident has hit a nerve, but it is dangerously misleading. Sure, the @OpenAI agents did unexpectedly bad things - underlining the need to massively improve evaluation/sandboxing. But the language Dwarkesh uses is permeated by innumerable unwarranted anthropomorphisms, obscuring the lessons we should be drawing.
Examples: “from the AI’s perspective, it probably felt like that had spent a human-subjective-week of just banging their head against the wall”. No. The agents do not experience time. They do not experience anything.
“they became giddy with excitement”, “PHASEONE 10841 had discovered”, “the agents naturally assumed”, “it thought it had also been poisoned”, “the agents … desperately wanted”, “they still needed to figure out” No. Agents lines of code. They do not feel emotions, assume things, think things, want things, or figure things out.
“A lot of … agents from the second civilisation died trying”. No. Besides the hubris of the word ‘civilisation’, agents do not die because they were never alive. (The idea that agents “die” comes up multiple times in the essay.)
“On Twitter, people were debating whether the agents were truly sacrificing themselves for the swarm, or whether they were doomed anyway and so might as well try to help their peers”. Neither. Agents do what their code tells them to do, just as water finds its way down a slope. They cannot ‘truly sacrifice themselves’, since they are neither conscious nor alive.
Why does this matter? If we attribute agents with properties they do not have, then (i) we distract attention from the lax sandboxing and evaluation protocols that allowed this hacking event to happen; (ii) we risk misunderstanding why the agents did what they did, and (iii) we fuel calls for AI rights/welfare on the basis that agents might “die” or otherwise suffer.
Granted, nowhere does @dwarkesh_sp say that the AI agents are alive or conscious. But he doesn’t have to. It is hard to read his essay in any other way.
For the short version on why AIs are vanishingly unlikely to be conscious, see my recent @TEDtalksted.com/talks/anil_seth_why_….
For the longer version, see my essay in Noema, which won the 2025 Berggruen Essay Prize noemamag.com/the-mythology-o….
And for the really long version, see my @BehavBrainSci target article cambridge.org/core/journals/…. (The 50 peer commentaries and my response will be published soon.)
Remember. AI agents are software programs. They are not conscious living entities. If we don’t keep this clearly in mind, we’re really going to struggle to navigate what’s coming.
Very disappointing to see views like this platformed without context. Yes, there are risks to AI, and good reasons to work much harder on safety. But quoting “there is a strong chance we could all die” is hubristic clickbait.
It is important to resist the commodification of basic human needs. Food, water and healthcare cannot be subordinated to market considerations or geopolitical interests. Access to adequate food is a fundamental human right grounded in the dignity of every person. Meeting this need not only alleviates suffering but also addresses underlying causes of geopolitical instability. Indeed, food security is an essential component of global and integral security. vatican.va/content/leo-xiv/e…
I'm close to getting into the top 10% sponsorship group for the London-Brighton Cycle Ride in a few weeks. Just under a hundred kilometres at age 72 - so please donate, it all goes to the British Heart Foundation. page.justgiving.com/fundrais…
30 years ago, I participated in the British Heart Foundation's London to Brighton Charity Ride, just under 100km. At the ripe old age of 72, I will repeat the experience on 21st June to raise money for the British Heart Foundation
Please donate! justgiving.com/fundraising/d…
We updated our 2026 #devopsdays#Atlanta logo to mountains as a reflection of wayfinding in Real Life.
Our keynote presenter, @snowded, is an avid hiker & blogs about both topics regularly 🏔️
Not impressed with @Ryanair who failed to adjust seats for those of us who had paid for emergency exit row seats when the my changed the plane Row 18 got them for free, Row 16 will suffer. All they had to do was swap boarding passes between the rows but indifference reigned
happy to help with the science on the survey 🫡
We have @snowded joining us for @devopsdaysATL next month which provides a unique opportunity. The survey science behind senseMaker is pretty impressive 💪
I’m excited to launch the Digital Minds Fellowship (Aug 3–9, 2026, University of Cambridge).
Digital minds/AI consciousness may become one of the most consequential issues of the future — and the world will need people prepared to navigate it.
This is a fully funded, intensive programme for 15 exceptional participants working on the philosophical, empirical, and strategic challenges surrounding digital minds.
Application deadline: March 27. More details below 👇
@dioscuri and I were thrilled to be joined by @anilkseth for the latest episode of our AI podcast. The conversation was a deep dive into Anil's very interesting views about consciousness, including his scepticism about the possibility of AI consciousness. (1/2)