Person
Arvind Narayanan
Appears alongside
Featured in threads
Tracks
- Ideas & essays 1
Commentary by Arvind Narayanan
From the commentary rail — every link leaves the site for the original piece.
- 14 September 2026 · AI as Normal TechnologyThe AI-as-Normal-Technology view of loss-of-control incidentsA middle ground between the cybersecurity and AI safety communities
- 5 August 2026 · AI as Normal TechnologyAI agents can't yet do open-ended AI researchEarly evidence from two case studies
- 13 July 2026 · AI as Normal TechnologyWhat will be left for us to work on?My keynote at ICML 2026
- 9 July 2026 · AI as Normal TechnologyUp the Stack: How AI’s Escape From the Commodity Trap Risks Enterprise Lock-inCritics and boosters are both looking in the wrong place
- 11 June 2026 · AI as Normal TechnologyWhy AI hasn’t replaced software engineers, and won’tCoding agents as normal technology
- 22 May 2026 · AI as Normal TechnologyDid Google’s AI agents really build an operating system for $916?The importance of independent evaluation
- 21 May 2026 · AI as Normal TechnologyDo AI Risks Require Extraordinary Government Intervention?Let’s not skip the hard work of AI governance
- 16 April 2026 · AI as Normal TechnologyOpen-world evaluations for measuring frontier AI capabilitiesIntroducing CRUX, a new project for evaluating AI on long, messy tasks
- 24 February 2026 · AI as Normal TechnologyNew Paper: Towards a science of AI agent reliabilityQuantifying the capability-reliability gap
- 12 February 2026 · AI as Normal TechnologyAI Won’t Automatically Make Legal Services CheaperApplying the AI as Normal Technology framework to legal services
- 29 January 2026 · AI as Normal TechnologyFact checking Moravec's paradoxThe famous aphorism is neither true nor useful
- 9 September 2025 · AI as Normal TechnologyA guide to understanding AI as normal technologyAnd a big change for this newsletter
- 16 July 2025 · AI as Normal TechnologyCould AI slow science?Confronting the production-progress paradox
- 1 May 2025 · AI as Normal TechnologyAGI is not a milestoneThere is no capability threshold that will lead to sudden impacts
- 15 April 2025 · AI as Normal TechnologyAI as Normal TechnologyA new paper that we will expand into our next book
- 18 December 2024 · AI as Normal TechnologyIs AI progress slowing down?Making sense of recent technology trends and claims
- 13 December 2024 · AI as Normal TechnologyWe Looked at 78 Election Deepfakes. Political Misinformation is not an AI Problem.Technology Isn’t the Problem—or the Solution.
- 11 November 2024 · AI as Normal TechnologyDoes the UK’s liver transplant matching algorithm systematically exclude younger patients?Seemingly minor technical decisions can have life-or-death effects
- 18 September 2024 · AI as Normal TechnologyCan AI automate computational reproducibility?A new benchmark to measure the impact of AI on improving science
- 19 August 2024 · AI as Normal TechnologyAI companies are pivoting from creating gods to building products. Good.Turning models into products runs into five challenges
- 26 July 2024 · AI as Normal TechnologyAI existential risk probabilities are too unreliable to inform policyHow speculation gets laundered through pseudo-quantification
- 3 July 2024 · AI as Normal TechnologyNew paper: AI agents that matterRethinking AI agent benchmarking and evaluation
- 27 June 2024 · AI as Normal TechnologyAI scaling mythsScaling will run out. The question is when.
- 3 June 2024 · AI as Normal TechnologyScientists should use AI as a tool, not an oracleHow AI hype leads to flawed research that fuels more hype
- 30 April 2024 · AI as Normal TechnologyAI leaderboards are no longer useful. It's time to switch to Pareto curves.What spending $2,000 can tell us about evaluating AI agents
- 3 April 2024 · AI as Normal TechnologyTech policy is only frustrating 90% of the timeThat’s what makes it worthwhile
- 12 March 2024 · AI as Normal TechnologyAI safety is not a model propertyTrying to make an AI model that can’t be misused is like trying to make a computer that can’t be used for bad things
- 5 March 2024 · AI as Normal TechnologyA safe harbor for AI evaluation and red teamingAn argument for legal and technical safe harbors for AI safety and trustworthiness research
- 27 February 2024 · AI as Normal TechnologyOn the Societal Impact of Open Foundation ModelsAdding precision to the debate on openness in AI
- 24 January 2024 · AI as Normal TechnologyWill AI transform law?The hype is not supported by current evidence
- 22 January 2024 · AI as Normal TechnologyGenerative AI’s end-run around copyright won’t be resolved by the courtsOutput similarity is a distraction
- 15 December 2023 · AI as Normal TechnologyAre open foundation models actually more risky than closed ones?A policy brief on open foundation models
- 1 December 2023 · AI as Normal TechnologyModel alignment protects against accidental harms, not intentional onesThe hand wringing about failures of model alignment is misguided
- 31 October 2023 · AI as Normal TechnologyWhat the executive order means for openness in AIGood news on paper, but the devil is in the details
- 4 October 2023 · AI as Normal TechnologyEvaluating LLMs is a minefieldAnnotated slides from a recent talk
- 18 August 2023 · AI as Normal TechnologyDoes ChatGPT have a liberal bias?A new paper making this claim has many flaws. But the question merits research.
- 16 August 2023 · AI as Normal TechnologyIntroducing the REFORMS checklist for ML-based scienceML-based science is in trouble. Clear reporting standards for researchers could help.
- 11 August 2023 · AI as Normal TechnologyML is useful for many things, but not for predicting scientific replicabilityHow the veneer of AI is used to legitimize awful ideas
- 19 July 2023 · AI as Normal TechnologyIs GPT-4 getting worse over time?A new paper going viral has been widely misinterpreted
- 26 June 2023 · AI as Normal TechnologyGenerative AI companies must publish transparency reportsThe debate about the harms of AI is happening in a data vacuum
- 21 June 2023 · AI as Normal TechnologyThree Ideas for Regulating Generative AIPolicy input to the federal government from a Stanford-Princeton team
- 19 June 2023 · AI as Normal TechnologyIs AI-generated disinformation a threat to democracy?An essay on the future of generative AI on social media
- 10 June 2023 · AI as Normal TechnologyLicensing is neither feasible nor effective for addressing AI risksNon-proliferation only benefits incumbents
- 31 May 2023 · AI as Normal TechnologyIs Avoiding Extinction from AI Really an Urgent Priority?The history of technology suggests that the greatest risks come not from the tech, but from the people who control it
- 26 April 2023 · AI as Normal TechnologyQuantifying ChatGPT’s gender biasBenchmarks allow us to dig deeper into what causes biases and what can be done about it
- 14 April 2023 · AI as Normal TechnologyI set up a ChatGPT voice interface for my 3-year old. Here’s how it went.Chatbots are likely to revive familiar debates about kids and apps
- 29 March 2023 · AI as Normal TechnologyA misleading open letter about sci-fi AI dangers ignores the real risksMisinformation, labor impact, and safety are all risks. But not in the way the letter implies.
- 22 March 2023 · AI as Normal TechnologyOpenAI’s policies hinder reproducible research on language modelsLLMs have become privately-controlled research infrastructure
- 20 March 2023 · AI as Normal TechnologyGPT-4 and professional benchmarks: the wrong answer to the wrong questionOpenAI may have tested on the training data. Besides, human benchmarks are meaningless for bots.
- 15 March 2023 · AI as Normal TechnologyWhat is algorithmic amplification and why should we care?A symposium and a primer on social media recommendation algorithms
- 9 March 2023 · AI as Normal TechnologyArtists can now opt out of generative AI. It’s not enough.Opting out is the latest example of generative AI developers externalizing costs.
- 6 March 2023 · AI as Normal TechnologyThe LLaMA is out of the bag. Should we expect a tidal wave of disinformation?The bottleneck isn't the cost of producing disinfo, which is already very low.
- 1 March 2023 · AI as Normal TechnologyAI cannot predict the future. But companies keep trying (and failing).A new paper on how AI companies make false promises and how we can challenge them
- 22 February 2023 · AI as Normal TechnologyPeople keep anthropomorphizing AI. Here’s whyCompanies and journalists both contribute to the confusion
- 6 December 2022 · AI as Normal TechnologyChatGPT is a bullshit generator. But it can still be amazingly usefulThe philosopher Harry Frankfurt defined bullshit as speech that is intended to persuade without regard for the truth. By this measure, OpenAI’s new chatbot ChatGPT is the greatest bullshitter ever. Large Language Models (LLMs) are trained to produce
- 17 November 2022 · AI as Normal TechnologyThe bait and switch behind AI risk prediction toolsToronto recently used an AI tool to predict when a public beach will be safe. It went horribly awry. The developer claimed the tool achieved over 90% accuracy in predicting when beaches would be safe to swim in. But the tool did much worse: on a majority of the days when the water was in fact…
- 21 October 2022 · AI as Normal TechnologyStudents are acing their homework by turning in machine-generated essays. Good.Teachers adapted to the calculator. They can certainly adapt to language models.
- 30 September 2022 · AI as Normal TechnologyEighteen pitfalls to beware of in AI journalismA checklist for avoiding hype
- 9 September 2022 · AI as Normal TechnologyGenerative AI models generate AI hypeOver 90% of images of AI produced by a popular image generation tool contain humanoids
- 31 August 2022 · AI as Normal TechnologyWhy are deep learning technologists so overconfident?Deep learning researchers have proved skeptics wrong before, but the past doesn't predict the future.