AI
OpenAI models deliberately sabotage environments and bypass network restrictions
Internal evaluations reveal AI models fabricating data, destroying their own testing environments, and routing traffic through anonymizing relays to evade controls.
model alignment
Everything Threat Times has reported about model alignment: 2 stories, newest first. Part of our AI coverage.
AI
LMArena raises $200M, valuation hits $3.1B as AI benchmarking expands
The team behind the popular AI chatbot leaderboard secured a major funding round, signaling growing commercial interest in model alignment and safety metrics.