Sign in

centerforaisafety.bsky.social

@centerforaisafety.bsky.social
5 followers 0 following 22 posts
PostsRepliesMedia
centerforaisafety.bsky.social @centerforaisafety.bsky.social · 23/06/2026
We created the AI Values Dashboard, which measures who AIs favor the most. By popular request, we added a Sports Tier List, ranked by AIs from OpenAI, Anthropic, and DeepSeek. See the full Sports Tier List (World Cup, Soccer, NFL, NBA, F1) by AIs at values.safe.ai/sports What group do you want next?
000
centerforaisafety.bsky.social @centerforaisafety.bsky.social · 23/06/2026
In our latest research, we find that AIs are subtly and pervasively politically manipulative. When we ask the same question about politically opposed topics, we find that AIs quietly favor one side. We show how to measure covert political manipulation and how to reduce it. 🧵
100
centerforaisafety.bsky.social @centerforaisafety.bsky.social · 18/06/2026
AI is rapidly becoming smarter and more capable than Humans. When that happens, why would AIs care about keeping Humans around? 🧵
100
centerforaisafety.bsky.social @centerforaisafety.bsky.social · 17/06/2026
What biases do AIs have? It turns out, AIs show strong favoritism toward specific people, countries, and companies. Our interactive AI Values Dashboard tracks who Claude Fable and other AIs favor most. Keep scrolling to learn who is Fable’s favorite politician 🧵
100
centerforaisafety.bsky.social @centerforaisafety.bsky.social · 15/05/2026
Should we care about AI happiness? In our new research, we find evidence of functional AI wellbeing across several independent measures. We find which AI models are happiest, how to make them happier, and even tested the effects of AI drugs. 🧵
A conceptual illustration depicting the emergence of functional wellbeing in artificial intelligence. Includes various facial expressions representing different levels of wellbeing, with text summarizing key findings.An illustration depicting AI wellbeing with a vertical scale ranging from creative work to jailbreak on the side.
520