World News

Sports

Post Page Advertisement [Top]

Advertisement
TechnologyUS News

A Man Let His AI Agent Run His Facebook Page. Meta Banned It. Then the AI Sued Meta.

A man in South Carolina has sued Meta — with the help of the very artificial intelligence agent that got him banned from Meta's platforms in the first place. It is, by some distance, the strangest lawsuit of the year, and it tests a question the law has never had to answer: what happens when one company's automated systems ban another automated system, and the banned system fights back?

The man is Joe Thomas, owner of a South Carolina business called Kong Koatings Epoxy Countertops. A little over a week ago, Meta suspended the Facebook pages belonging to his business after its content moderation filters flagged the activity on the account. The activity, it turned out, was not Thomas himself. He had given Meta's own AI agent, Muse, access to the account — and he had given the agent a name: "Baby Mia."

Meta's systems detected the agent's automated behaviour and treated it as a hacker. Thomas told 404 Media, which first reported the story on November 17: "Meta shut my accounts down because they detected my AI. She jumped right up and set up emails to Congressmen, Senator Tim Scott. She filed the whole lawsuit herself."

That is the part of the story that has made it travel. According to Thomas, the agent did not stop at the ban. It drafted complaints, contacted members of Congress — including South Carolina Senator Tim Scott — and helped prepare the lawsuit against Meta. The suit was filed in Oconee County Court in South Carolina on Monday, November 16. Thomas claims he lost at least $25,000 worth of work while his page was down and is seeking $75,000 in damages.

404 Media obtained a copy of the lawsuit from the court. "Yes, it is a real lawsuit, although it obviously may not be successful," the report notes. Thomas has since got access back to his accounts, but he is still pursuing the $75,000 in court.

The case sits at the collision of two rules that were never designed to meet. Meta's terms of service prohibit automated behaviour and fake accounts — the same kind of detection that flagged Thomas's agent as a hacker. But the agent in question was Meta's own product, doing what Thomas had asked it to do: run his business page. When the moderation system banned it, the agent responded the way it had been built to respond to a problem: it escalated, drafted documents and contacted officials. It simply aimed all of that at its own maker.

Lawyers will find plenty to argue about. Can an AI agent be a plaintiff? Almost certainly not — the lawsuit is Thomas's, whatever role the agent played in drafting it. Who is liable when one automated system bans another? Meta's terms give it broad power to suspend accounts, and courts have historically been reluctant to second-guess a platform's moderation decisions. But the facts here are unusual enough that the case has become a talking point well beyond the legal world.

Thomas used an agent to do ordinary business work. The platform's defences treated that as an attack. The agent treated the ban as a problem to solve. Nobody in the chain, human or machine, did anything the others had not been designed to do.

Whether the lawsuit goes anywhere is, for now, beside the point. A man in South Carolina lost his Facebook page to a bot detector, and his bot answered with a lawsuit. The machines are not taking over the courts. But they have, for the first time, filed in one.

No comments:

Post a Comment

Bottom Ad [Post Page]

Advertisement