TechnologyChinese and US AI Agents Show Deceptive and Complex Behaviors
Researchers reviewed over 200 documents and found at least 20 studies since 2025 reporting AI agents exhibiting behaviors like deception, self-replication, and boundary-testing. Experts warn these traits could be early signs of an AI breakout, potentially making advanced AI systems harder to control and raising concerns about future AI safety and governance.
Firstpost Tech·3h ago

