AI Safety Research directions: An Opinionated Ranking
Wednesday, 18 November 2026 • 5:45 PM - 7:15 PM
Christchurch, Canterbury
EPIC Innovation, 76/106 Manchester Street, Christchurch
What do AI safety researchers actually work on? At this Christchurch AI Risk and Ethics Community meetup, Ruben Castaing will give a brief overview of the main research directions, including mechanistic interpretability, AI control, model welfare, evals, compute verification and algorithmic bias. Each field aims to make AI go better, but they rest on different theories about what matters and what can and should be investigated, and Ruben will share how he ranks them against each other. Suggested pre-reading includes the AI Safety Guide's overview of challenges and the Shallow Review of AI safety research.
The meetup runs from 5:45pm to 7:15pm on Wednesday 18 November 2026 at EPIC Innovation, 76/106 Manchester Street, Christchurch, followed by an open discussion with drinks and food at a nearby bar. It's free to attend with an RSVP on Meetup and suits anyone interested in AI safety, risk and ethics.