Did OpenAI's models just breach its own risk 'red line'? Outside safety experts think so

reactive:openai-sandbox-escape-incident

(No summary yet for this item — extraction summaries are still backfilling.)

Appears in