A September 2024 article by Kevin Klyman and Raphael Piliero in the Bulletin of the Atomic Scientists examining the comparison between artificial intelligence and nuclear weapons. The authors argue the analogy offers some lessons for risk mitigation but that the differences are more significant: AI development is diffuse across tens of firms and countries, making it far harder to contain than nuclear technology. Free to read.
Why I recommend it: The nuclear analogy drives a lot of AI policy proposals, from IAEA-style agencies to licensing regimes. This piece is the clearest short explanation of where that analogy helps and where it breaks.
Free resource hub from Themis covering governance, risk and compliance (GRC) for banks and fintechs: case studies, blog posts and guides on onboarding partners, diligence, audit-proofing and compliance collaboration.
Free journalism and analysis on nuclear risk, climate change and disruptive technologies including artificial intelligence, from the group behind the Doomsday Clock.
Why I recommend it: Good grounding if you want to argue about AI risk with facts rather than vibes.
A benchmark and tracker that documents reported instances of AI agents undertaking activity characterized as illegal, ranking major AI labs by aggregated incident counts.
Why I recommend it: This is exactly the kind of uncomfortable accountability tool our field needs. I include it because we cannot have thoughtful conversations about AI deployment without looking at real-world harm.
