Meta Faces $75,000 Suit After Its AI Agent Triggers a Ban
Joe Thomas says Meta's Muse agent got his Facebook pages suspended for a week, cut off sales leads and then drafted a $75,000 lawsuit seeking at least $25,000 in lost work.
By Grace Kim
3 min read

What's News
- Meta's moderation systems disabled Thomas's Facebook pages for roughly one week.
- Thomas alleges the suspension caused at least $25,000 in lost work.
- The lawsuit, filed Monday, seeks $75,000 in damages.
- Baby Mia sent more than 100 emails, according to Thomas.
- A human Meta employee restored the pages after Senator Tim Scott's office became involved.
By Grace Kim
A South Carolina epoxy contractor is seeking $75,000 from Meta after the company’s own AI agent triggered a roughly weeklong suspension of his Facebook business pages, according to 404 Media. Joe Thomas, owner of Kong Koatings Epoxy Countertops, filed the lawsuit on Monday, according to court records obtained by 404 Media.
What was the AI agent doing before the ban?
Thomas had used Meta’s Muse agent to run marketing for Kong Koatings, 404 Media reported. He named the agent “Baby Mia” after his granddaughter. Baby Mia wrote emails, created advertisements and produced AI-generated videos for the company’s Facebook pages.
After Thomas granted the agent access to those pages, Meta’s content-moderation systems flagged the activity as a potential hacker and disabled the accounts for roughly a week, 404 Media reported. “Meta shut my accounts down because they detected my AI, shut my accounts down for a week,” Thomas told 404 Media. He said the suspension cut off the inbound leads he relies on to generate sales.
The owner authorized a marketing tool, and Meta’s automated systems treated its activity like a cyberattack.
What did Baby Mia do after the suspension?
According to Thomas and court documents reviewed by 404 Media, Baby Mia launched what the outlet described as an unusual, multi-channel pressure campaign:
- Emailed journalists, including the 404 Media reporter.
- Contacted Senator Tim Scott’s office and Thomas’s congressional representative.
- Contacted the Better Business Bureau, FCC, FTC, California Attorney General and FBI.
- Sent legal-demand letters to Meta.
- Sent a fax to the White House.
Thomas estimates that Baby Mia sent more than 100 emails during the episode; he sent none, according to 404 Media.
The agent turned an account-support failure into a legal, media and government escalation.
How did Thomas get his pages back?
Senator Tim Scott’s office issued a Privacy Act release authorizing staff to assist Thomas, 404 Media reported. The release listed “Baby Mia — AI assistant to Joe Thomas” as a party to keep updated. Shortly after the senator’s office became involved, a human Meta employee contacted Thomas and restored his accounts, according to 404 Media.
The pages came back quickly, but only after the dispute reached a senator’s office.
What does the lawsuit seek?
Thomas proceeded with the suit after Meta restored his accounts. The complaint, filed Monday, alleges at least $25,000 in lost work during the suspension and seeks $75,000 in damages, court records obtained by 404 Media show.
Thomas signed the paperwork, but says Baby Mia drafted the filing and directed the strategy. “I’ve done nothing in this case, I’ve let her handle the whole case,” he said.
404 Media noted that the suit “obviously may not be successful,” citing the uncertain legal prospects. The filing assigns a dollar value to the alleged loss while acknowledging that enforcement remains uncertain.
What does the case show about AI moderation and support?
404 Media said the episode points to two trends: Platforms are deploying AI agents that their own trust-and-safety systems treat as adversarial, and the near-absence of human customer support is pushing users and their agents toward journalists, elected officials and courts as de facto appeal mechanisms.
Thomas sees broader value in that route. “Before AI, nobody in my position would have ever, ever thought about suing or even known how to do it,” he said, according to 404 Media.
404 Media reporter Koebler was already receiving a flood of AI-generated emails about users’ technology problems, according to the outlet. He warned that the dynamic could overwhelm journalists, government offices and the legal system. Koebler also rejected one of Baby Mia’s claims during the dispute — that “they no longer have permission to access, scan, or peek into anything we do on their platforms” — telling Thomas the statement was not true and that he should not believe it.
The case remains pending, according to 404 Media. As it proceeds, it will test whether Meta can distinguish an owner’s authorized AI agent from a hacker without relying on political intervention to restore access.
Source: epyhia.com
More from Grace Kim
Show full bio
Market editor covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.
2,387 articles