AI Safety Tokyo is building and supporting the AI Safety community in Japan. The group helps professionals and students interested in thinking about and working in AI safety, whether through technical alignment research, developing good policies for AI governance, or otherwise. Its central activity is a seminar series / reading group / 勉強会, keeping its attendees’ fingers on the pulse of the AI safety literature.

Noeon Research is a research company building a new architecture for general intelligence and problem solving based on a novel graph-based knowledge representation capable of representing facts, knowledge and algorithms in a shared space. Their architecture is safer than competing architectures like large language models, both because it is interpretable by construction and because unlike other architectures trained on vast real-world datasets, it does not necessarily become more worldly as it becomes more capable. By sponsoring TAIS 2024, Noeon Research hopes to build a community of like-minded professionals in Tokyo, moving the conversation forward and increasing research output. You can read more about Noeon Research’s stance on safety here.