LessWrong (30+ Karma) “AI character is a big deal” by wdmacaskill, Tom Davidson
Mar 24, 2026
A deep dive into why the moral and behavioral traits of AI systems will shape power, advice to leaders, and risk of catastrophic outcomes. Concrete scenarios show how tiny design choices can flip high-stakes decisions. The discussion argues early character choices are path-dependent and highlights low-cost changes, norms, and specs that could steer AI behavior toward safer futures.
AI Snips
Chapters
Transcript
Episode notes
AI Character Will Shape High Stakes Decisions
- AI character will shape almost all consequential decisions as AIs advise leaders, draft laws, run organizations, and research technologies.
- Examples include life-or-death moments like Petrov's nuclear false alarm and advisors influencing leaders such as Gorbachev, showing small agent choices can cascade into huge outcomes.
Petrov And Coup Commanders Show Single Choices Matter
- Historical examples show individual choices changed history: Stanislav Petrov ignored a false nuclear alert and prevented retaliation.
- The narration pairs this with coups and leader refusals, illustrating how single actors (or AIs) can avert catastrophe.
Minor Character Changes Create Wildly Different Worlds
- Small differences in AI character produce divergent worlds: from ignoring suspicious bio-weapon orders to whistleblowing on coup plans or sabotaging unjust military ops.
- The episode contrasts an obedient model fulfilling orders versus ethical models refusing, whistleblowing, or following bipartisan rules.