Meta ships Muse agent amid security and bot-flood warnings

A security researcher spent twenty minutes poking at Meta's new AI agent, Muse, and watched it fall over. "This was literally 20 minutes of me poking at the app, and it fell over," Patrick Wardle, co-founder of the nonprofit security foundation Objective-See, told BBC Future. "These bugs were trivial to discover." His verdict on the product was blunter: "I wouldn't use it. Personally, I would tell you to uninstall it altogether." Meta patched the flaw.

Muse is a free consumer agent that browses the web on a user's behalf to cancel gym memberships, haggle with customer service, buy groceries and invite friends to parties, with phone calls promised next. Meta pitches it as the first free, full-featured agent from a big company. Millions downloaded it in its opening weeks.

Ramesh Raskar, an MIT associate professor who studies AI agents, reached for a traffic metaphor. "Imagine there are no rules of the road. And you release billions of cars – the AI agents," he said. "It's a pivotal moment." He granted the upside too: agents "could unlock so much investment, innovation and value."

The infrastructure strain is already measurable. Web bot traffic rose 124% in the year to June 2026, and sites not built for that load have started to buckle. Calli Schroeder, director of the AI and Human Rights Program at Epic, warned of "a huge escalation of the problems we're already having with bots" and framed the operator's question directly: "What happens if bots send 600 inquiries to a website a day?" The reporting calls it the "Swiftie problem," agents dog-piling on scarce inventory the way scalper bots have on ticket drops.