Arvind Narayanan
Computer Scientist
Princeton computer scientist and co-author of “AI Snake Oil” (2024) and “AI as Normal Technology” (2025), which frame AI as a real, useful, general-purpose tool — neither miracle nor menace. He debunks the hype economy while using the tools himself: a skeptic of salesmanship, not of the science.
In his words
AI snake oil is AI that doesn’t work as advertised and probably can’t work as advertised.
The kinds of risks that we see in the news are exotic. But the kinds of harms that are already widespread — that are happening to people every day — those you don’t hear about.
Biography
Arvind Narayanan (born 1983) is a professor of computer science at Princeton and director of its Center for Information Technology Policy. Earlier in his career he demonstrated how supposedly anonymized datasets could be re-identified, and he has become one of the most cited academic skeptics of Silicon Valley’s AI marketing — though skeptic, in his case, does not mean opponent.
AI Snake Oil — and “normal technology”
With his student Sayash Kapoor, Narayanan turned a popular newsletter into the 2024 book “AI Snake Oil.” Its argument is precise: he draws a sharp line between generative AI, which he calls “a genuinely new and interesting technology,” and predictive AI — the systems used to screen job applicants, set bail or flag welfare fraud — which he says frequently cannot do what vendors claim. The harms he emphasizes are mundane and pervasive, not science-fictional.
His 2025 essay “AI as Normal Technology” makes the balance explicit: AI is transformative in the way electricity and the internet were transformative — powerful, worth adopting, but a tool that humans integrate into workflows rather than an autonomous oracle or a separate species. It is a deliberate rejection of both utopian and doom framings, and it recommends using AI, not resisting it.
Something like ChatGPT has very little in common with the predictive AI that banks might use to calculate someone’s credit score. These are two very, very different technologies.
Narayanan is also a critic of existential-risk framing, which he sees as distracting from the everyday failures of deployed systems — a stance that puts him at odds with figures like Eliezer Yudkowsky. Yet he readily praises the tools he finds useful, from spellcheck to code assistants, and resists being cast as anti-AI. His target is the hype economy, not the science.
Where they stand in the war
Voices for AI
Sources & further reading
Canonical record: https://battlelines.ai/topic/arvind-narayanan