AI agents abound, unbound by rules or safety disclosures
MIT CSAIL's 2025 AI Agent Index puts opaque automated systems under the microscope AI agents are becoming more common and more capable, without consensus or standards on how they should behave, say academic researchers.... [...]