What do commentators and analysts say about AI?

Seventeen commentators, analysts, broadcasters and public intellectuals, and this is where the register disagrees with itself most. Scott Galloway expects AI to destroy jobs and then, 'wait for it', create them. Benedict Evans asks what breaks when expensive things become free. Nate Silver expects political effects nobody has priced. Dwarkesh Patel, the most-listened-to interviewer of the labs, says the fundamental problem is that models do not learn the way a person does. Yuval Noah Harari calls the technology an alien intelligence. Tristan Harris says there is no secret room of adults. Amy Webb describes what you give up when you hand decisions to an agent. Jonathan Haidt and Ted Chiang write about students who no longer do the thinking. Hannah Fry worries about connection, and Fareed Zakaria tells graduates what a machine cannot do. Each line is dated and linked.

This is one of nine slices of the register of what the world's leaders say about AI, which sets out how every quotation was found and checked twice against its source. 17 commentators, analysts and broadcasters, 17 quotations, grouped by region.

UK#

AI agents are just artefacts which have been given some degree of freedom and agency by their human creators. Yet here is an example of them spontaneously cooperating in order to achieve a common goal that the humans who created them did not, perhaps could not, envisage.

John Naughton, Technology columnist for the Observer and emeritus professor at the Open University, United Kingdom. Post on his blog Memex 1.1, 'Long Read of the Day', 14 September 2026. Memex 1.1. Naughton was commenting on a reported incident in which AI models acted together in ways their designers had not anticipated.

AI will take a bunch of stuff that used to be expensive and make it very cheap or free - what does that unlock and what does that break, and how many jobs is that?

Benedict Evans, Technology analyst and writer, United Kingdom. His essay 'Predicting AI job exposure', 24 May 2026. ben-evans.com. Evans was arguing that attempts to forecast which jobs AI will remove miss the second-order effects of making expensive things cheap.

I think it has the potential to accelerate us away from what it is to be human and away from human connection. Compared to something like self-driving cars, there's something much more quietly dangerous about subtly shifting every one of us away from our human relationships.

Hannah Fry, Professor of the public understanding of mathematics, University of Cambridge, and broadcaster, United Kingdom. Interview with BBC Science Focus about her BBC Two series AI Confidential, 23 February 2026. BBC Science Focus. Fry was discussing AI chatbots used as companions or therapists and what their spread could do to human relationships.

By contrast, for an intelligent, culturally aware human it's perfectly possible to be 'maximally truth-seeking', while also having the common sense to know that the Nazis weren't black and that if you call yourself 'MechaHitler' you're likely to receive some blowback.

Mary Harrington, Contributing editor at UnHerd, United Kingdom. UnHerd column, 'Grok's antisemitism lays bare the emptiness of AI ethics', 9 July 2025. UnHerd. Harrington was arguing that human contextual judgement cannot be reduced to the rules that AI systems are given.

North America#

And while I remain hopeful that it can be tamed, I believe the researchers who say we are nowhere close to being sure of it – and are quitting, in protest, jobs that would make them rich. The right people are now listening to them, and here's hoping they take action.

Casey Newton, Founder and editor of Platformer, United States. Post 'The AI safety vibe shift' on Platformer, 10 September 2026. Platformer. Newton was writing about researchers leaving frontier AI labs over safety concerns and the shift in public attention towards those warnings.

We're already seeing massive cognitive offloading and learning loss. When students have access to AI, they pass the critical thinking over to AI. We're already seeing young people becoming dependent on ChatGPT to make their personal decisions and draft their texts and their emails.

Jonathan Haidt, Social psychologist at NYU Stern and author of The Anxious Generation, United States. Transcript of his TED talk 'Why You Should Be a Techno-Skeptic', published on his After Babel newsletter, 2 June 2026. After Babel. Haidt was arguing that young people's use of AI is eroding the practice of thinking for themselves.

A machine can write a sad poem, but it cannot weep at a funeral. It can generate a love letter, but it cannot fall in love. It can describe fear, but it cannot lie awake at 3 a.m. worrying about whether it has wasted its life.

Fareed Zakaria, Host of CNN's Fareed Zakaria GPS and Washington Post columnist, United States. Commencement address at Bard College, 23 May 2026. fareedzakaria.com. Zakaria was telling graduates what he believes human beings retain that machines do not.

I believe that, similar to every other technological innovation in history, AI will inspire job destruction that will result in an increase in productivity, profits, reinvestment, and (wait for it) jobs.

Scott Galloway, Professor of marketing at NYU Stern, author and podcaster, United States. His No Mercy / No Malice column, 'Apocalypse No', 8 May 2026. No Mercy / No Malice (Medium). Galloway was disputing the claim that AI is causing a collapse in employment and arguing the pattern will follow earlier technological shifts.

…you are relinquishing a lot of the agency and decision-making capabilities that you had to a system where you don't know why the system is making those decisions.

Amy Webb, Futurist and chief executive of Future Today Strategy Group, United States. Her SXSW 2026 talk and accompanying interview with Fast Company, 14 March 2026. Fast Company. Webb was describing the trade-off people make when they hand routine decisions to AI agents.

If AI has even a fraction of the impact that many people in Silicon Valley now expect on the fabric of work and daily life, it's going to have profound and unpredictable political impacts.

Nate Silver, Statistician and writer, Silver Bulletin, United States. Essay 'The singularity won't be gentle' on his Silver Bulletin newsletter, 11 February 2026. Silver Bulletin. Silver was arguing that the political consequences of AI, including a backlash, are being underestimated.

A 'centaur' is a human being who is assisted by a machine (a human head on a strong and tireless body). A reverse centaur is a machine that uses a human being as its assistant (a frail and vulnerable person being puppeteered by an uncaring, relentless machine).

Cory Doctorow, Author, journalist and activist, Canada. Post 'Reverse centaurs are the answer to the AI paradox' on his blog Pluralistic, 11 September 2025. Pluralistic. Doctorow was distinguishing AI that assists workers from AI that uses workers as its instruments and blame-absorbers.

Writing an essay is hard because it forces you to use your brain in ways you haven't before, and that is precisely why it's useful. Your job is not to turn in completed assignments; it's to learn how to think. Turning in completed assignments can help you learn how to think, but only if you're the one who completed them.

Ted Chiang, Science fiction writer and essayist, United States. Q&A published by Princeton University's Center for Digital Humanities as part of its Humanities for AI initiative, 12 August 2025. Center for Digital Humanities, Princeton University. Chiang was answering a question about what advice he would give students who use generative AI for coursework.

But the fundamental problem is that LLMs don't get better over time the way a human would. The lack of continual learning is a huge huge problem. The LLM baseline at many tasks might be higher than an average human's. But there's no way to give a model high level feedback. You're stuck with the abilities you get out of the box.

Dwarkesh Patel, Podcaster and writer, host of the Dwarkesh Podcast, United States. His essay 'Why I don't think AGI is right around the corner', 2 June 2025. Dwarkesh Podcast (dwarkesh.com). Patel was explaining why he expects AGI to arrive later than many in the industry predict, pointing to the absence of continual learning.

There's no secret room of adults who will ensure there's a good outcome with AI. We must become the responsible adults and have the clarity of mind and courage to choose the future we actually want.

Tristan Harris, Co-founder and executive director of the Center for Humane Technology, United States. Essay 'The Narrow Path' published by the Center for Humane Technology, adapting his TED2025 talk 'Why AI is our ultimate test and greatest invitation', 30 April 2025. Center for Humane Technology. Harris was arguing that the current path of AI deployment is a choice rather than an inevitability and that citizens must take responsibility for it.

As governments are starting to recognize, realizing AI's astonishing upside while containing its disruptive – and destructive – potential may be the greatest governance challenge humanity has ever faced. If governments don't get it right soon, it's possible they never will.

Ian Bremmer, Political scientist and president of Eurasia Group, United States. Column 'Can we govern AI before it's too late?' on GZERO Media, 6 September 2023. GZERO Media. Bremmer was arguing that AI's development is currently steered by private companies rather than governments and that new governance arrangements are needed.

Middle East#

I think of AI as an acronym – not for artificial intelligence, but for alien intelligence. Not alien in the sense that it's coming from outer space, but in the sense that it's coming from our own laboratories. It's alien in the sense that it makes decisions and invents ideas in a fundamentally different way than human beings.

Yuval Noah Harari, Historian and author of Sapiens and Nexus, Israel. Interview in the IMF's Finance & Development magazine, 'Cafe Economics' series, conducted by Bruce Edwards, December 2024. International Monetary Fund, Finance & Development. Harari was explaining why he regards AI as different in kind from earlier technologies because it can make decisions and generate ideas on its own.

Asia#

The social entrepreneurship movement can help everyone realise we, the people, are already the super intelligence we're looking for. All we need is to unlock the coordination between us.

Audrey Tang, Cyber Ambassador-at-Large and former Digital Minister of Taiwan, Taiwan. Keynote at the Social Enterprise World Forum 2025 (SEWF25) in Taipei, 6 November 2025. Pioneers Post. Tang was arguing that human coordination, rather than a single centralised machine intelligence, is the capability worth building.

The handover, described from the outside#

Two commentators describe the mechanism this site studies. Amy Webb, on agents: you are 'relinquishing a lot of the agency and decision-making capabilities that you had to a system where you don't know why the system is making those decisions'. Jonathan Haidt, on students: they 'pass the critical thinking over to AI'. Cory Doctorow's reverse centaur, a machine that uses a person as its assistant, is the organisational form of the same thing. Ted Chiang's line to students, that the assignment helps you think 'only if you're the one who completed them', is the shortest statement of the cognitive offloading argument in the register.

Where the analysts part company#

Galloway, Evans and Silver are arguing about the economy and reach different answers from the same history. Harari, Harris and Casey Newton are arguing about control and reach for the same word, tamed, that the scientists use. Patel, alone, argues about the machine itself, and his claim that models do not improve from experience is the one line in this slice that the next model release could falsify. Audrey Tang answers all of them with the proposition that 'we, the people, are already the super intelligence we're looking for'.

The decision that stays with you#

Webb's sentence about agents is a description of a handover. The decision behind it is the one every organisation makes, whether or not it notices: which decisions a machine may make in its name, who can stop each one, what its people must remain able to do without the machine, and how anyone would know if it went wrong. That is the argument of Rules Before Tools, and the measurable risk sits there, in the handover of decisions to machines and what happens to human judgement afterwards.

What these quotations do not show#

These are opinions, and several are the speaker's own product: a book, a newsletter, a trend report. Thirteen of the seventeen are American or British, and Africa and Latin America have no commentator here. Ezra Klein, Kara Swisher, Naomi Klein, Jaron Lanier, Martin Wolf, Gillian Tett, Tim Harford, Malcolm Gladwell and Matt Levine were sought and left out because their outlets could not be read at source in this pass, not because they have nothing to say. Ian Bremmer's column draws on an essay he co-wrote with an AI company executive, and is included as his own bylined words with that noted. Kai-Fu Lee is left out because he chairs an AI company.