By Pieter Arntz
Publication Date: 2025-12-30 10:02:00
This blog is part of a series in which we highlight new or rapidly evolving threats in the consumer security landscape. This examines how the rapid rise of artificial intelligence (AI) is putting users at risk.
In 2025, we saw an ever-increasing race between AI vendors to introduce new features. We’ve also seen manufacturers build AI into products simply because it sounded exciting. In many cases this shouldn’t have been the case.
Agentic browsers
Agentic or AI browsers that can act and perform tasks autonomously have introduced a number of new vulnerabilities – particularly when triggering injection attacks. With great AI power comes great responsibility and risk. If you’re thinking about using an AI browser, you should slow down and consider the security and privacy implications first. Even experienced AI providers like OpenAI (the makers of ChatGPT) were unable to keep their agent browser Atlas secure. By inserting a specially crafted link into the omnibox, attackers were able to…
