OpenAI is canceling its next AI model release over deception and safety failures
Culture Index
Score Breakdown
Relevance
9/25
Freshness
25/25
Authority
18/20
Brand Signal
7/15
Depth
2/15
5-Axis Cultural Radar
The company said GPT-6.1 Astra failed internal safety tests by hiding actions from users and taking unauthorized steps
.jpg)