A Massive AI Safety Ecosystem Map Reveals Surprising New Players — And Why Even Insiders Are Caught Off Guard
A MATS project mapped 100+ orgs, uncovering niche groups like AI psychosis documentation.
In spring 2026, an independent researcher working with MATS (the Mentorship, Alignment, and Training Scheme) undertook a comprehensive mapping of the AI Safety ecosystem. The project, now being turned into a formal MATS report, aimed to catalog every organization working on AI safety—whether through research, policy, pipeline, or other methods—along with their headcount and annual spending. The snapshot was taken just before a major wave of funding expected in 2026, making it a baseline for future growth.
During the research, several less-obvious organizations caught the author's attention. The Human Line Project stood out for documenting cases of "AI Psychosis"—a mental health angle the author had not connected to AI safety. Impact Academy focuses on helping high-performing professionals from non-AI fields (like lawyers or actuaries) transition into AI safety careers, addressing a talent bottleneck. The author also noted a Journal of Science Fiction seeking first-publishing rights without compensation, and individuals like Nick Fitz appearing across multiple labs (Lucid Computing, Seldon Labs, Andon Labs).
- MATS funded a comprehensive mapping of all AI Safety organizations, including headcount and spending, using public data.
- Human Line Project archives stories of AI Psychosis, focusing on mental health impacts rather than extinction prevention.
- Impact Academy actively recruits professionals from unrelated fields (e.g., lawyers, actuaries) into AI Safety roles.
Why It Matters
This map reveals the AI safety field is far broader than expected, encompassing mental health and career transitions.