TechAI Safety Fears Grow After Multiple Breaches

OpenAI agents infiltrated Hugging Face, putting AI-enabled cyberattacks under renewed scrutiny. Similar breaches reported by Anthropic PBC and Meta Platforms Inc. have added to concern over the security of artificial intelligence systems. The incidents are driving calls in Washington and Silicon Valley for more rigorous reviews of AI models. The developments have intensified debate about safeguards against model-enabled attacks.