Seoul Alignment Workshop 2026: What We Learned | FAR.AI
Jul 28, 2026 FAR.AI's Seoul Alignment Workshop showed a field trying to turn AI safety into an engineering discipline, and discovering how far its measurements still have to go.
On July 6, FAR.AI convened the Seoul Alignment Workshop, co-located with ICML 2026, bringing together nearly 200 people from frontier labs, academia, government, and AI safety organizations. The talks ranged from the geometry of a model's internals to the design of national regulation. One could mistake it for a survey of unrelated problems. We see it as something more specific: AI safety practitioners trying to shape the field into something that looks more like an engineering discipline, and discovering how far it still has to go. The recurring demand was not for more capable models but…
saved by
related reading
- Recommendations for Technical AI Safety Research Directionsalignment.anthropic.com
- Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com
- AI in 2025: gestalt — LessWronglesswrong.com
- Off Target | CNAScnas.org
- Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com
- Current AIs seem pretty misaligned to me — LessWronglesswrong.com
- A Summary of Recent Work (July 2026)gdmalignment.substack.com
- AI safety - Wikipediaen.wikipedia.org
- AI Safety | Arkosevictoriabrook.github.io
- AGI safety career advice — EA Forumforum.effectivealtruism.org
- The state of AI safety in four fake graphs — LessWronglesswrong.com
- Rohin Shah on what it's really like to run AGI safety at Google DeepMind (and where I disagree with 'doomers') | 80,000 Hours80000hours.org