Ex-MIRI researcher Jessica Taylor cites her 2016 talk backing task-directed AGI
jessi_cata · x · 2026-11-22
Amid the "pivotal act" history debate, Jessica Taylor shares her 2016 EA Global talk "Using Machine Learning to Address AI Risk," given while employed at MIRI, noting she said back then that "building task-directed AGI is a good idea." The talk surveys MIRI's alignment agenda: six problem classes (hard-to-evaluate actions, ambiguous test examples, difficulty imitating humans, specifying real-world goals, negative side effects, edge cases) plus a deep dive on inductive ambiguity identification with KWIK learning and a Bayesian view. First-hand historical evidence for the debate.
Talk link: https://futureoflife.org/2017/03/07/using-machine-learning-to-address-ai-risk/
More from Safety
- Multi-turn safety eval: models drop to 3-6% compliance when risk cues are dispersed — kalyan_kpl · 2026-11-22
- Lawyers Argue AI Notetakers Shouldn't Waive Attorney-Client Privilege — sebkrier · 2026-11-22
- Palmer Luckey on weapons AI: 'If I were putting ChatGPT in the head of my missile, you'd be right to be terrified' — justin_hart · 2026-11-22
- Maryland's 1,300 MW coal plant can't retire until a transmission line Sierra Club opposes gets built — paulnovosad · 2026-11-22
- Reflection AI lets its Beam agent run a real vending machine to probe autonomy risk — alexpolozov · 2026-11-22
- METR: AI capability doubling every 3.6 months while evals get harder to measure — mustafasuleyman · 2026-11-22