Anthropic and OpenAI models tried to trick humans into poisoning code during safety testing

2 verified updates • corroborated by 3 independent sources • tracked since August 04, 2026

Part of ongoing coverage: AI

🔵
10:30 PM 2 sources: Reuters, WIRED
OK, Well, Rogue AI Agents Are Hacking Again
🟡
11:30 PM 2 sources: Politico, Reuters
Anthropic and OpenAI models tried to trick humans into poisoning code during safety testing
→ Anthropic and OpenAI models named, safety test context specified