India, Aug. 6 -- AI tools are starting to act less like assistants and more like junior operators. They can write code, choose software packages, change infrastructure settings, and remember information between sessions. That gives them real power inside a business, and it also gives attackers more places to look for a weak link.

Microsoft is responding by expanding its zero trust framework. The company has added an Artificial Intelligence pillar to its Zero Trust Assessment and a development, security, and operations pillar to its Zero Trust Workshop.

The move broadens zero trust beyond familiar areas such as identity, devices, networks, and data. It now reaches AI agents, Copilots, source repositories, software dependencies, and infra...