OpenAI and DeepMind researchers warned AI labs are racing "blindfolded" toward self-improving systems
- DeepMind: 4 events in the last 90 days
- Previous: 13 days earlier · AI must not outrun safety controls, DeepMind co-founder warns
What happened
Current and former employees said labs reward model-builders over those urging caution, as debate over AI safety becomes a global political issue
Summary assembled by rule from the sources below