The deterrent effect of gathering DNA from people arrested for felonies is large, and illustrates something important about crime. 1/2
The paper I’ve been most obsessed with lately is finally out: https://www.nbcnews.com/tech/tech-news/ai-changing-style-substance-human-writing-study-finds-rcna263789! Check out this beautiful plot: it shows how much LLMs distort human writing when making edits, compared to how humans would revise the same content.
We take a dataset of human-written essays from 2021, before the release of ChatGPT. We compare how people revise draft v1 -> v2 given expert feedback, with how an LLM revises the same v1 given the same feedback. This enables a counterfactual comparison: how much does the LLM alter the essay compared to what the human was originally intending to write? We find LLMs consistently induce massive distortions, even changing the actual meaning and conclusions argued for.
This is second time we've used Anthropic Interviewer and the first time we've deployed it at scale. Quite accidentally, we ended up conducting (what we believe is) the largest qualitative study ever
I'm a mixed-methods social scientist by training. Traditionally, when it came to understanding what people think, that meant quantitative analysis of lower resolution data (polls, surveys, etc.) or hand-wavey analysis of in-depth qualitative data. Using Claude to conduct *and* analyze interviews bridges that tradeoff between breadth and depth
AI also makes access much, much easier. Had we run this study in person, in the real world, it would have taken hundreds (if not several thousand) enumerators many 1000s of hours to conduct. It also affords us access to places we could otherwise never go. I once led a five-person team in Tanzania that reached a few hundred people. It took 3 weeks. In this study we heard from people 80,000 people in 159 countries, in cities and rural areas, in daily life and in war zones, and more, in just one
I'm still, even after months, beginning to wrap my head around the scale of this work. Like, to a social scientist, it's quite unbelievable. This could produce dozens of dissertations! It is also, of course, imperfect—certainly speaking to an AI is different than speaking to a person—and as a team we're all still figuring out how to make this research as useful as possible: what questions to ask and how, what to analyze and why, and how that all feeds back into what we do as a company. This is, as we say in the blog, a brand new form of social science
Hat tip to @saffronhuang for leading this for the past few months. Here's one of my favorite quotes
Excited to be on Odd Lots to talk about the politics of AI.
AI today is less important than it will ever be.
Over the past year, AI rose in issue importance faster than any issue we track — it's now more important to voters than climate change, child care, and abortion.
Today in @TheArgumentMag, Milan Singh and I do a deep dive into the data on who is ordering food delivery.
https://www.theargumentmag.com/p/whos-really-ordering-all-that-doordash
What that viral Anthropic jobs chart really means, by @jburnmurdoch and @sarahoconnor_
• Task vs. Job Automation: The chart distinguishes between theoretical AI exposure and actual real-world usage. It emphasizes that LLMs typically automate specific tasks rather than entire occupations; if AI handles low-value tasks while humans focus on high-value ones, the technology may augment and enhance a profession rather than displace it.
• Subjectivity of the Data: The metrics used are inherently "fuzzy." The theoretical exposure is based on subjective assessments by researchers and GPT-4, which often ignore regulatory, legal, and organizational hurdles. Meanwhile, Anthropic’s real-world data reflects Claude's usage logs, which don't always clearly distinguish between AI helping a worker (augmentation) and AI replacing a workflow (automation).
• The "Rorschach Test" Effect: Despite high theoretical exposure in fields like computing, there is currently no evidence in labor market data of increased unemployment in those sectors. Consequently, the chart acts as an inkblot test: optimists see it as "hype-busting" proof of AI's limited reach, while pessimists see the vast gap between theoretical and actual use as a sign of massive disruption yet to come.
https://www.ft.com/content/2cd79c2c-d1f0-4417-a06f-2572868ee858 vía @ft
Alarming story from @georginaquach:
At one UK university, the scramble to attract lucrative international students to the new London campus saw thousands admitted without the necessary English or academic skills, widespread use of ghostwriters, and fraudulent attendance logging
R to @jburnmurdoch: Arguably the most eye-opening stat:
After tightening its recruitment processes, the campus only took on 31 students this year, down from 1,624 (a fall of 98%).
Full piece: https://www.ft.com/content/aa845112-58a6-48c5-9278-35e809465607
This week's column: choose your university wisely
Post-1992 providers have been rapidly expanding business, law and computing courses. Yet the returns for students 5 years after graduating from these courses have, to date, been woeful
1/4
New newsletter: THREE REASONS TO BE A PARENT
Britain’s politicians have pushed through a number of well-intentioned policies that have been ultimately disastrous.
Another great column by @jburnmurdoch
The AI Shift is the @FT’s latest newsletter, from the fabulous team of @sarahoconnor_ and @jburnmurdoch
Do we really know which jobs are most at risk from AI?
https://as.ft.com/r/c1791cf8-710c-48c9-95ef-8b8d114e6b79
The graduate premium is falling in the UK but rising in the US.
However even if the grad premium was rising everywhere, it still might have nothing to do what students learn at university. It could be the result of the "signal" sent by university attendance - a signal that could be replaced by something far cheaper and less time-consuming.
Even in America, where the grad premium is rising, there are plenty of people who would argue that it's still the result of a signal - not the result of the human capital gained from being at university.
There is a huge academic debate about this which is very hard to settle one way or the other, partly because we have such poor assessment data on universities.
I don't think the value of university is 100% signal. I think it's a mix of human capital & signals.
If we want to tilt it more towards human capital, we need better assessment data. We need to know which universities are the best at teaching certain skills (and which skills are most valuable in the job market).
Graduate earnings data cannot tell you this.
I just don't think it's good enough for the graduate premium to be the sole measure of university impact. You wouldn't measure a hospital by the later earnings of the people it treats! We shouldn't measure universities that way either!
https://substack.nomoremarking.com/p/does-a-university-education-help
British Growth and the Graduate Premium: Britain's growth problems will not be solved by more graduates. https://tomforth.co.uk/graduatepremium/
R to @jburnmurdoch: Full piece here: https://www.ft.com/content/649d3c64-b8e5-4979-9f0c-9aebd43642e2
R to @jburnmurdoch: Just one more example of how the UK is a formerly dynamic and growing economy behaving as if it is still a dynamic and growing economy, and then being surprised when things that used to work (and work for other countries that haven’t stagnated) no longer work here.
R to @jburnmurdoch: Fundamental point: the graduate earnings premium is not really measuring anything to do with higher education or graduates.
It’s measuring “how well-paid are your high-earners (which tends to be where your graduates are) relative to lower-earners?”
R to @jburnmurdoch: Which place has the highest grad wage premium?
The US, because it has lots of super well-paid professional jobs and quite high earnings inequality.
Which places have the lowest grad wage premiums? Sweden and Denmark (~25% each), because they have much lower earnings inequality.
R to @jburnmurdoch: For the UK today, think Italy of the 1990s and 2000s.
Just as back then the ranks of under-employed and emigrating Italian graduates were an indictment of the Italian economy and labour market rather than the Italian higher education system, the same is true of the UK today.
R to @jburnmurdoch: 2) Sustained increases in the UK’s minimum wage squeeze the graduate earnings premium from the other side.
• Graduate earnings (shaped more by economic dynamism) have been eroded
• Non-graduate earnings (shaped more by policy) have risen
• Grad premium doubly squeezed
R to @jburnmurdoch: 3) Weak economy means the UK has simply not created as many top professional jobs as peer countries (roughly half the growth in highly skilled occupations).
The result is a growing share of UK grads working in non-grad jobs.
R to @jburnmurdoch: I’ve shown another lens on this same thing previously: the UK’s top income decile has had a very rough couple of decades (and the top half as a whole has done much worse than the bottom half).
Who works in top-paying jobs? Graduates. That’s the erosion of the premium right there
R to @jburnmurdoch: 1) The UK economy’s growth and dynamism have never recovered from the financial crisis. This productivity crash has had a scarring effect on earnings at the top of the market — the jobs most graduates work in.
The real sign British education has failed is the number of people responding to this chart with "that’s what happens when too many people go to university"
HE has expanded in all of these countries, and in every one apart from UK that didn’t erode the graduate earnings premium.
R to @jburnmurdoch: “But it expanded more here than elsewhere”.
Not really relevant. Why did US and Netherlands going from 25% college grads to 40% college grads see their earnings premiums rise, while the same in the UK led the premium to fall?
The answers are quite straightforward:
https://www.ucl.ac.uk/mathematical-physical-sciences/events/2026/feb/rdr-celebrating-ucl200-communicating-risk-uncertain-world Looking forward to speaking with @jburnmurdoch at this event next Tuesday in London
Whereas the graduate premium has increased in most rich countries, it has plummeted in Britain since 1997.
Earnings for British graduates have shrunk (next pic).
->
Can feed algorithms shape what people think about politics? Our paper "The Political Effects of X's Feed Algorithm" is out today in @Nature and answers "Yes".
https://www.nature.com/articles/s41586-026-10098-2
"If you had come to me ten years ago and told me about the leading AI models that we have today, that GPT 5.2 and Claude Opus 4.6 would be publicly available via a cheap API call, I would have thought that we’d be seeing something like mass unemployment."
Why don't European companies innovate? It is common to blame expensive energy, high taxes, anti-growth politicians, interest groups, and green regulations.
But California has the same problems, and has created the world's most innovative companies.
Europe's problem is labor law. Compared with America, it's far harder to let workers go when a business doesn't work out.
https://worksinprogress.co/issue/why-europe-doesnt-have-a-tesla/
- It costs a large company roughly four times more to fire a worker in Germany or France than the US.
- German law requires employers to consider age, years of service, family obligations, and disability status when deciding who to lay off. Employees who would be least impacted by losing their job are prioritized for dismissal.
- German employees who take on a caregiving role are fully protected from dismissal for two years from the date they begin caregiving.
- Factory closures in Germany regularly lead to payments of over €200,000 per employee.
- French companies must be prepared to show a court that their financial results are struggling enough to make layoffs necessary.
- To avoid the difficulties of formal dismissals, many European companies entice workers to depart voluntarily, with payouts of up to four years' salary.
Taken together, a German worker is ten times less likely to be fired in a given year than an American worker. This high cost of firing makes failures more expensive. It pushes big European companies away from taking risks and leads them to concentrate on safe, unchanging areas.
Europe has the ingredients needed to succeed. Its citizens are educated and inventive; it has excellent infrastructure and the rule of law; and its culture is not that different from the one it had fifty years ago, when its companies were world-beating. If Europe wants to a Tesla or a Google, it only needs to make it cheaper for companies to fail. My new piece for @WorksInProgMag.
Anthropic should move to London. Or, at least, dual-list in London, with a significant presence here.
Here's why:
1. Anthropic is spiritually British. Their philosopher-in-residence, Amanda Askell, is Scottish, and Jack Clark (cofounder) is English, as are many other staff. Askell would be further away from Elon Musk in London.
2. Unlike other US labs, Anthropic cares more about safety, risks and good regulation. Compared to the EU, Britain's AI regulation is more focused on safety (rather than "ethical AI") and growth-oriented.
3. It's not good for the world that all the frontier labs (excl. DeepMind, sort of) are US-based, and therefore subject to the whims and potential control of whoever is in the White House. If AGI happens, do you want Trump controlling it?
4. Britain desperately needs a stake in the post-AI economy. AI could replace much of our services sector could be at risk from AI. Our energy is too expensive for data centres or manufacturing. But we have talent, AI expertise (e.g. AISI) and global reach. The solution? Give British people a capital stake in frontier labs.
5. Anthropic wants to expand its AI for science work, where the UK is a global leader. As part of the Oxford-Cambridge growth corridor, HMG should co-fund the world's biggest AI for science campus near London, with Anthropic as the anchor funder & tenant.
The UK government should do everything it can to get Anthropic to move. We are spending £500m on Sovereign AI, to support "national champions". Instead, the British state could be an early shareholder in a newly-London listed Anthropic.
Did you notice £2.3bn giveaway in Budget to small group of public sector pensioners? Another transfer from young to old.
After years of companies pouring hundreds of billions into DB pensions the companies, not pensioners, should benefit from surpluses.
https://www.thetimes.com/article/0b38edfa-9e14-4e5f-8cab-23e0e5dc192d?shareToken=80ba1bd667fae211132fe270263786d0