Dear AI Companies: Stop the “Doom Trolling” | AI Reality Check

What It Covers

Cal Newport, computer scientist and author, coins the term "doom trolling" to describe how AI companies like OpenAI and Anthropic simultaneously predict catastrophic outcomes from their own products — including 25% extinction odds and 50% white-collar job automation — while continuing to raise capital and accelerate development.

Key Questions Answered

• •Doom Trolling Defined: AI companies engage in a communication pattern Newport calls "doom trolling" — publicly predicting their own products could cause economic collapse or human extinction, then continuing product development and IPO preparation unchanged. Anthropic released its "terrifying" Mythos model just six weeks after alarming world leaders about its dangers, suggesting performative fear over genuine concern.
• •The Two-Option Moral Test: When AI companies predict catastrophic harm, only two explanations exist: they genuinely believe it, in which case halting development is the only ethical response, or they are manufacturing fear to inflate valuations. Newport argues both options are morally indefensible — one is negligence, the other is monetizing public anxiety for early shareholders.
• •Financial Incentive Behind the Fear: AI companies benefit financially from doom rhetoric because existential-scale narratives justify trillion-dollar valuations that their actual business models cannot support. OpenAI is essentially a money-losing natural language search engine; Anthropic is primarily a developer utility. Apocalyptic framing converts them into meme stocks worthy of speculative investment capital.
• •Ignore Future-Tense AI Claims: Newport's practical filter: disregard any statement from AI companies framed in the future tense. Evaluate only current products on their present utility and cost. Scary news articles about AI typically reflect AI company doom trolling directly, not independent journalistic analysis, so major publication coverage carries no additional credibility.
• •Burden of Proof Inversion: When resisting AI doom narratives, individuals feel pressured to disprove catastrophic claims rather than requiring companies to prove them. Newport reframes this: remarkable claims require remarkable evidence, and it is the doomer's responsibility to substantiate predictions, not the skeptic's job to refute each new benchmark or white paper animation.