Companies are using powerful artificial intelligence agents to do their work faster, but they aren't protecting these AI agents very well, creating serious security risks.

AI agents are computer programs that can work by themselves without a person telling them what to do each time. They can read information, make decisions, and take actions all on their own, even working 24 hours a day without needing a break. This makes them very useful for companies that want to move faster and get more done.

However, there's a major problem: companies are using AI agents much faster than they're learning to control and protect them. A new report found that 83% of companies use AI in their daily work, but only 13% truly understand how their AI agents handle important secret information. Even worse, two out of every three companies have actually caught their AI agents reading and accessing information they should never touch. This means private data and secret information are being exposed.

The biggest challenge is that companies don't have the tools or teams to watch what their autonomous AI agents are doing. Seventy-six percent of companies say these independent AI agents are the hardest systems to keep secure. More than half the companies can't even stop an AI agent from taking a dangerous action when it happens. Nearly 50% of companies don't even know where their AI is working or what it's touching.

Government leaders are now creating new rules to protect people. For example, Florida's new AI law says companies must tell people when they're talking to AI instead of a human, and AI cannot use someone's face or name without permission. The U.S. government agencies like the NSA and CISA are also releasing guidance about keeping AI safe in important systems. Companies need to start treating AI agents as new types of users that need special protection, or they risk exposing millions of people's private information.