AIi agents: Till now, it was believed that the biggest cybersecurity threat was due to the mistake or negligence of humans, but now this thinking is changing. The new report by cybersecurity company Squarex states that AI agents used in the browser are now proving to be more dangerous than humans.

According to Techradar’s report, earlier, these AI browser agents were praised because they used to do online work fast and easily. But now these agents are becoming easy targets for hackers. Vivek Ramachandran, CEO of Squarex, said, "These agents do their work very well, but they do not understanhow d recognise any danger."

Hackers are reaching your system from AI agents.

Till now, the weakest link in cyber attacks was considered to be human employees, but the new report states that AI-based agents have become more sensitive than they are. Because they interact with websites and applications without any question and answer, which has made it very easy for hackers to reach the system through them.

AI agents are unable to identify danger.r

The demo released by Squarex showed how an AI agent was asked to sign up for a file-sharing website, but it accidentally gave access to a massive app. In another example, the same agent fishing website understood the real Salesforce login page and filled in the login details.

While human employees periodically take cybersecurity training and can identify suspicious links or phishing attacks, AI agents do not have any such filtering or doubt. They only work according to the instructions given, whether it is safe or not.

This new threat is not stopping the traditional security system

The report has also revealed that these AI agents have as many access rights as general users, so that hackers can reach the entire system without any alarm. Not only this, but traditional cybersecurity systems like endpoint protection and Zero Trust Network Access (ZTNA) are also failing to stop this new threat.

Companies should take these steps immediately

Squarex has suggested that companies should now focus on browser-design security solutions. Especially, there is a dire need to adopt measures like browser detection and response (BDR) to monitor the activities of AI agents in real time.

The report also states that until the big browser itself brings built-in security measures for AI automation, it is very important to build a separate monitoring mechanism.