Cohere adds new security controls to its AI agent platform

Canadian artificial intelligence developer Cohere announced on Monday that it has updated its North agent harness to enhance security and operational control. This latest iteration, known as North 2, introduces advanced features including skills, libraries, and automations designed to streamline how autonomous agents interact with existing software and data repositories. The update specifically focuses on improving access control through strict protocols, managing token consumption, and facilitating better collaboration across enterprise platforms.

These enhancements aim to address growing corporate concerns regarding the security and unpredictability of autonomous AI agents. By implementing standardized operating procedures through its new "skills" feature, Cohere allows businesses to share successful task workflows, which promotes consistency and prevents redundant work. For organizations, this development may necessitate a review of current business software to determine if updates or new connection methods are required to integrate safely with these more controlled, sovereign AI frameworks.