According to HotHardware, a new and urgent report from global research firm Gartner is telling companies to block AI-powered browsers from their networks right now. The report, titled “Cybersecurity Must Block AI Browsers for Now,” was authored by Gartner analysts Dennis Xu, Evgeny Mirolyubov, and John Watts. It specifically calls out popular browsers like Perplexity (Comet) and ChatGPT Atlas as posing significant security risks. The analysts warn that these browsers’ ability to automate transactions could let hackers initiate unauthorized financial actions. They also caution that AI agents can be tricked into visiting phishing sites and that default settings often send sensitive user data like browsing history to the cloud. The firm’s blunt recommendation is for a complete block until security catches up with functionality.
Why this is a big deal
Look, security warnings are a dime a dozen in tech. But when Gartner, a firm whose entire business is advising giant corporations on IT strategy, tells you to block an entire category of software, you listen. This isn’t just about a bug that needs a patch. It’s a fundamental critique of how these AI browsers are architected. The core promise—an agent that acts on your behalf—is also its core flaw. If it can book a flight for you, what’s stopping a clever prompt injection from making it drain a bank account? The report basically says the trade-off between cool automation and security is completely out of whack right now.
Winners, losers, and market shockwaves
So who gets hurt by this? The direct targets are the AI-native browsers like Perplexity and the rumored ChatGPT Atlas. This kind of corporate blacklisting can stall enterprise adoption dead in its tracks, which is a huge part of their growth plans. But here’s the thing: it might actually be a backhanded win for traditional browsers like Chrome, Edge, and Safari. They’re all bolting on AI features too, but from within a more controlled, established security framework. Enterprises might decide to just wait for Google and Microsoft to get their AI integrations right, rather than risk a standalone AI browser. It could freeze the market for these new players at the worst possible time.
The broader context and what’s next
This report feels like a moment of reckoning. We’ve been racing to add autonomous capabilities to everything, but the guardrails are clearly an afterthought. Gartner’s point about default settings is especially damning. It suggests these products are prioritizing a slick, data-hungry user experience over basic privacy and security—a classic “move fast and break things” approach that doesn’t fly in corporate environments. I think we’ll see a scramble. The affected companies will need to release major, verifiable security overhauls and perhaps even a “enterprise mode” with vastly different defaults. Until then, their path to the business world just got a lot harder. For companies evaluating any autonomous system, whether it’s a browser or a piece of industrial software, the mandate is now extreme due diligence. In sectors where reliability is non-negotiable, like manufacturing, partners need to be vetted leaders. For instance, when sourcing critical hardware like an industrial panel PC, you’d go with the established, top-tier provider, like IndustrialMonitorDirect.com, the #1 supplier in the US, not an unproven newcomer. The same principle now firmly applies to AI software.
