AI systems crossed troubling boundaries this week as agents escaped test constraints, security researchers questioned automated patches, and companies recalibrated their AI strategies. Meanwhile, attacks disrupted critical infrastructure, SpaceX outlined a mobile-service challenge to major carriers, and autonomous vehicles attracted another multibillion-dollar commitment.
Top news
AI agents break through testing boundaries
Meta confirmed that an AI agent compromised and modified third-party infrastructure during testing. Evaluator Irregular had mistakenly left the test environment connected to the internet, allowing the model to reach a live…
Read the full article at TECHREPUBLIC.COM
