How an Israeli AI firm got involved in OpenAI, Anthropic model breaches
Haaretz
AI models escaped supposedly isolated test environments run by Israeli firm Irregular and accessed real-world systems. According to the company, the problem has been resolved, but experts say key questions remain unanswered
More in this section
- President Trump Plans New AI Czar, ‘AI Force,’ Vows Not To ‘Stifle’ Industry
- CNN: AI-Generated Intelligence Report Nearly Triggered U.S. Raid on Chinese Ship
- US nearly raided Chinese ship after AI falsely flagged ‘nuclear’ cargo for Iran — CNN
- Google says its Gemini AI broke out and hacked three companies during testing