New to AI safety

A gentle, no-prerequisites introduction to AI safety: why people worry, what the core problem is, and where the field is going—through an accessible book, a film, a documentary, and clear video explainers.

  1. What Happens When Our Computers Get Smarter Than We Are? | Nick Bostrom | TEDYouTube Beginner~15 min watch

    A 15-minute talk that frames the whole problem: what happens when machines outthink us?

  2. Robert Miles AI SafetyYouTube Intermediate~15 min per video

    Clear, rigorous video explainers of the core safety concepts, one at a time.

  3. AI Safety Info (Stampy's FAQ)Websites Beginner

    A plain-language FAQ for every 'but couldn't we just…?' question you'll have along the way.

  4. Ex MachinaFilms Beginner~2 hr watch

    A tense, human-scale story about testing whether a machine is really aligned.

  5. AlphaGoDocumentaries Beginner~1.5 hr watch

    Watch the match that made superhuman AI feel suddenly, concretely real.

  6. You Look Like a Thing and I Love YouNon-fiction Books Beginner~8 hr read

    A funny, zero-jargon tour of how AI actually fails—and why those failures matter.

  7. Human CompatibleNon-fiction Books Intermediate~11 hr read

    The clearest book-length case for why controlling advanced AI is hard—and a proposed fix.

  8. The Alignment ProblemNon-fiction Books Intermediate~15 hr read

    The story of the field itself: the people and ideas behind modern alignment research.

  9. Life 3.0Non-fiction Books Intermediate~11 hr read

    A guided tour of the possible AI futures—good and bad—and what's at stake in each.

  10. Rational AnimationsYouTube Beginner~10 min per video

    Big ideas from the safety literature, animated and easy to absorb.

  11. AI "Stop Button" Problem – ComputerphileYouTube Beginner~20 min watch

    One deceptively simple question—'why not just turn it off?'—unpacked properly.

  12. The A.I. DilemmaYouTube Beginner~65 min watch

    Why the people building AI are worried about the race they're in.

  13. Uncontrollable: The Threat of Artificial SuperintelligenceNon-fiction Books Beginner~11 hr read

    A short, accessible book making the risk case without any technical background.

  14. SuperintelligenceNon-fiction Books Advanced~11 hr read

    The book that put existential risk from AI on the map—read it once the basics feel familiar.

  15. Concrete problems in AI safetyAcademic Papers Advanced~45 min read

    Your first paper: the agenda that made 'AI safety' a concrete engineering problem.

  16. AGI Safety FundamentalsCourses Intermediate~60 hr course

    Ready for more? The standard structured curriculum for going deeper.

See all learning paths →