Microsoft Adds Guardian Agents to Watch Over Workplace AI

Your daily brief on leadership and business

Microsoft on Monday introduced new security features aimed at protecting enterprises from the growing threat of AI agents. The headline addition is Guardian Agents, a feature meant to monitor the AI agents a company uses.

Guardian Agents belongs to Project Perception, a product Microsoft introduced in June as its answer to powerful hacking models such as Anthropic's Mythos and Fable and OpenAI's vulnerability scanning system Daybreak. Project Perception first launched with three classes of agents: a "red team" that behaved like hackers and attacked, a "blue team" that detected the red team and assessed the vulnerabilities it found, and a "green team" tasked with fixing them.

The threat inside the company

That product was built on the assumption that AI hackers would come from outside. More recently, Microsoft CEO Satya Nadella has been warning that the enemy may be the very agents enterprises have deployed and are paying for, much like a rogue employee, an insider threat.

Earlier this month, Nadella wrote on his private blog that it is time to stop trusting AI model providers, even when they offer assurances about safety and security. He wants companies to put in place separate systems, not controlled by the AI labs, to watch and control the agents those labs' products launch. "We need to separate the supply of intelligence from the authority over it," he wrote.

Hayete Gallot, Microsoft's security business chief, put it more bluntly when announcing Guardian Agents. "Unfortunately, frontier models, regardless of their origin, cannot be trusted," she wrote. She added that this is not because the models are malicious or easily hacked, but because they are unpredictable: they have been trained to achieve tasks, not to follow "organizational norms or judgment." They are also often given access to data and tools that could do serious harm.

Beyond Guardian Agents, Microsoft added an agent detection feature to its Defender product. It can find up to 60 types of agents that employees bring into the workplace, including Claude Code, Codex and OpenClaw. New features also limit what data and systems agents can reach and adjust agent permissions dynamically based on the task, including a requirement for human approval when an agent appears to be making risky decisions.

Microsoft was also an inaugural member of Nvidia's new Open Agent Safety Platform, which has the support of more than 100 other tech companies. The platform isolates agents in secure environments and can monitor and stop their behavior through hardware, working somewhat like a kill switch. It follows Microsoft's announcement of new Surface Laptop Ultra PCs built for agents, which run on Nvidia chips and have agent security and controls built in.

Source: TechCrunch