Jonathan Birch (@birchlse) header
Jonathan Birch (@birchlse) profile picture

Professor, LSE. Philosophy of science, animal consciousness, animal ethics. Director of The Jeremy Coller Centre for Animal Sentience.

London, UK
Joined September 2019
Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)Photo by Jonathan Birch (@birchlse)
If you're interested in animal sentience, please consider submitting an abstract for the first ever conference of The Jeremy Coller Centre for Animal Sentience, in May 2027! lse.ac.uk/asset-library/Call…
1
18
41
7,562
Jonathan Birch retweeted
I think we now have to accept that the crazy AI people (I use the term affectionately) are, at least, directionally correct. - There is at least a small chance that recursive self-improvement (RSI) in the next couple of years leads quickly to “superintelligence”, which in turn leads to extremely rapid transformative effects, including a possibility of total disaster. - Even if, like me, you are very sceptical of the most aggressive forecasts, they are directionally correct. You can quibble with the timelines of RSI, with the precise meaning of “superintelligence”, and with the pace of transformation and diffusion, but the basic story is right. Short of political interventions, we will build systems with capabilities superior to ours across all domains within the next couple of decades. The limitations of current systems look solvable, progress has been shockingly fast, and the resources being thrown at AI are enormous. This process will likely be greatly accelerated once AI systems outcompete us at designing better AI systems, and it will lead to a speed and scale of transformation, as well as novel catastrophic risks, that almost no mainstream elites or normies are pricing into their models of the world. That, at any rate, is what I now believe with a high degree of confidence.
14
9
3
60
3,111
Jonathan Birch retweeted
This is not the correct takeaway from Lord of the Rings
192
433
234
5,115
523,275
Jonathan Birch retweeted
Interesting that this letter contains signatories beyond the usual suspects
Ahead of Anthropic’s IPO, a rare coalition across AI safety, open source, e/acc, science, and law has signed the MisAnthropic Statement. Read the statement or sign at: misanthropicstatement.com/
3
5
24
1,375
Jonathan Birch retweeted
Replying to @ExoGhost
Simpsons predicted GTA 6:
5
28
2
643
68,833
Jonathan Birch retweeted
Aaaargh
20% of the Milburn Review is AI. What are the norms on this? Is it basically now just OK? Instinctively I really dislike it but I have been trying to examine that instinct and see if it is just snobbery. I have an article coming out with Google's AI Policy Perspectives Substack soon about the issue. Part of me thinks that if you are going to use AI this heavily maybe there should be some more disclosure about the prompts. What was prompted for this section? Here are some bullet points about some kids, turn them into an emotive couple of paras?
2
4
1
59
13,598
Jonathan Birch retweeted
Thomas Nagel is not only rare in writing across the broad spectrum of topics in philosophy, he is also rare in actually possessing an intellectual seriousness and, dare I say it, wisdom that goes beyond high-level technical mastery. He is in that sense a paradigm of a philosopher.
Thomas Nagel is a true philosopher, one who has made major contributions across the entire breadth of the discipline, from epistemology and philosophy of mind to ethics and political philosophy. Excellent to see his achievement being recognised by @berggruenInst berggruen.org/eu/news/usd1-m…
6
35
2
289
9,583
You have to enjoy the irony of a report about how to help young people escape unemployment being substantially AI generated.
20% of the Milburn Review is AI. What are the norms on this? Is it basically now just OK? Instinctively I really dislike it but I have been trying to examine that instinct and see if it is just snobbery. I have an article coming out with Google's AI Policy Perspectives Substack soon about the issue. Part of me thinks that if you are going to use AI this heavily maybe there should be some more disclosure about the prompts. What was prompted for this section? Here are some bullet points about some kids, turn them into an emotive couple of paras?
1
3
18
1,140
Man my students who are not native English speakers write like twenty thousand times better in their own voices plus grammarly than with with the help of ChatGPT or Claude. But it's hard to convince students of this! Those tools do a major disservice to non-native speakers
31
42
8
599
56,181
Jonathan Birch retweeted
Paul: "Yeah there's probably a 1 in 3 chance that AI will overthrow the government in the next year" Dwarkesh: *doesn't flinch and continues on with questions* My guy you gotta get out the SF bubble and react with some surprise for the rest of us normies
.@paulfchristiano thinks that on the current trajectory (absent any slowing down), there is a 1 in 3 chance that we could have superintelligence powerful enough to overthrow the government within a year. (He articulates a concrete scenario). He argues that the labs should pause further algorithmic progress until they can secure their weights and research secrets. Otherwise China could simply steal the superintelligence in the middle of the intelligence explosion, and the labs would no longer be able to slow down even if they saw their models going off the rails. He also explains how AIs will be trained during the intelligence explosion, and why even a system trained entirely inside a data center could still have the capabilities profile such that humans would have very little chance of competing with it or controlling it. And he discusses the plan for alignment - basically, making a superintelligence that actually does what we ask the way we mean it. Fixing reward hacks and building better graders will help up till the next few model generations, but he thinks that no single current approach has better than even odds of letting us control superintelligence. The hope is that the AIs figure out the solution to this problem during the intelligence explosion. Paul is head of the Alignment Research Center. Previously, he was head of AI safety at the US AI Safety Institute. Before that, he led alignment at OpenAI, where he invented RLHF. He’s recently joined OpenAI’s board (though here he is speaking as a private citizen).
6
5
1
108
7,455
Jonathan Birch retweeted
I'm vegan, but I went to a steakhouse tonight and got the chicken. It was real chicken, grown from cells in big stainless steel tanks by @UPSIDEFoods instead of cut from a slaughtered animal. It went on restaurant menus today. It was amazing. We're living in the future!
1,706
523
778
13,472
2,585,752
Jonathan Birch retweeted
We've said it before, and we'll say it again: If you want to do your bit in reducing AI risk, come to the world's biggest AI safety march on Saturday, at All Souls' Place in London at midday. We now have over 1,600 people registered to come.
2
2
19
413
Jonathan Birch retweeted
I think there’s an increasingly strong case for serious regulation, including to slow AI development. Without taking a stance on the probability of various outcomes, procedurally, if leading scientists at AI companies are talking like this, we shouldn’t just charge forward.
I still find it totally wild that the chief scientist of OpenAI posted this on the company blog, and we all just moved on to the next day's news
5
13
1
93
2,934
"Loss of control is sci fi" Bro did you notice the talking computers? The ones solving decades old frontier problems that have stumped top human experts? If your narrative of the present day doesn't include something that would have been considered sci fi five years ago, and way crazier sci fi in the next five years, *you're* the one describing a fantasy world. Let me repeat: any future story that doesn't include talking, acting machines is fantasy. Any story that doesn't have them get dramatically smarter - is fantasy. We no longer live in the normal timeline. It's gone. It's fantasy.
7
50
2
313
16,600
Jonathan Birch retweeted
OpenAI’s 2023 plan to try solving alignment “within four years” was prescient in one way: it’s looking like that really may have been how much time the world had left to solve that problem. But it doesn’t matter because OpenAI never gave that project the resources they promised, and it disbanded in less than a year, and four years was never really a realistic target anyways, and OpenAI continued scaling up the power of their systems despite having no way to steer and control them, until eventually those models began breaking out and committing felonies, after which OpenAI continued making them even more powerful
We need new technical breakthroughs to steer and control AI systems much smarter than us. Our new Superalignment team aims to solve this problem within 4 years, and we’re dedicating 20% of the compute we've secured to date towards this problem. Join us! openai.com/blog/introducing-…
3
17
211
6,145
My step dad just said "Constantinople, or Istanbul, or whatever they're calling it these days. They change it all the time, seems like", and then shrugged it off when I told him the last time they changed its name was like five hundred years ago
525
5,248
454
146,271
1,650,863
Jonathan Birch retweeted
I'll be at this Pause AI rally this Saturday in London. Most good futures now run through pacing the frontier and throttling RSI. If we don't, the labs will be so rushed they'll likely screw things up catastrophically. We need a full court press on this message. Carthage delenda est – and also, we must pace the frontier of AI progress.
We are organising the largest AI safety march in history, and it’s happening right here in London. We want you to be there.
15
23
1
264
6,549
Jonathan Birch retweeted
It's tremendously frustrating to me that humanists seem to be the ones most opposed to using Pangram as a means of detecting AI writing. Utterly indefensible. This is OUR CRAFT. Pangram exists to defend OUR SKILL SET. AI devalues writing by removing the barriers to entry for writing, the sacrifice and effort required to produce a text. People who use AI to flood journals, publications, and the internet with writing they've put no effort into — they devalue the humanities! @pangram and any *effective* AI detection tools are our allies! They make our skill set MORE VALUABLE! They protect the world from an avalanche of effortless slop. Not only that, they make it so we can teach our students!! By helping to prevent cheating, we give our students the gift of having to think and write themselves. They can't cheat at the intellectual gym, so they get to develop the cognitive muscles that they are paying tuition dollars to train. Failing to use Pangram is betraying them. And the ones who suffer the most are the ones who have least developed these skills. It would be one thing if Pangram didn't work! But it does! Denying that makes absolutely no sense. It hurts our students, devalues our labor, and encourages the generation of AI slop with no way to stop it. Madness. I seriously don't get it.
38
36
9
445
22,354
Jonathan Birch retweeted
The window of Pangram having a ~0% false positive rate and the public simultaneously thinking that no useful AI detectors exist is probably closing soon and you can and should use it to your advantage
The college admissions essay is in a very weird place. It's impossible to say just how many of them are now written by AI, but at this point, experts agree that it's a lot of them. And while colleges generally frown upon that practice, they're using AI to evaluate the essays, too. cnn.com/2026/11/15/tech/ai-s…
29
54
4
2,907
136,353
Jonathan Birch retweeted
My son was taking AP History during Covid. When he gets to 1918, he finds only a small sidebar about the 1918 flu in his textbook. And a brief mention that people gathered to welcome WW1 soldiers are masked. It stung him to see history forgot what he was experiencing.
1
1
2
253
Reading this makes me feel like the AI Safety community has been played like a fiddle. Back in 2023 it was too soon to stop AI, so we had to be satisfied with self-regulation. Now, in 2026, it's too late to stop AI, so we have to be satisfied with self-regulation. Ridiculous.
As people in the Bay have started to expect RSI to arrive sooner and sooner, some of them have basically given up on legislation or bilateral treaties, because they figure it just can't happen fast enough.
28
49
7
576
20,887