Sask. content creator says she's frustrated after Facebook mistakenly took down 22,000-follower page

A Saskatchewan content creator says her business has been thrown into turmoil after Facebook unexpectedly deleted her page.

Ashlyn George runs the The Lost Girl's Guide to Finding the World, a travel brand. She says she’s been fighting since Thursday to have her 22,000-follower Facebook page restored after Meta suspended it for breaking its community standards on "impersonation."

"The impact is incredibly stressful," said George.

"I've spent 13 years sharing adventure stories, photos and writing with my followers, and to have it disappear overnight is both sad and frustrating, and disappointing knowing that it could be gone forever."

George says she’s kept her content consistent and wasn’t impersonating anybody. She doesn’t know why her page was flagged, though she’s dealt with harassment in the past and worries a user might be reporting her account.

"I feel a little bit helpless. I mean, there's an option to appeal, but sometimes the appeal button is there, sometimes it's not," she said. 

The takedown has turned into a days-long saga of dealing with Meta’s customer service program, which is dominated by AI service agents, said George.

Her account was reactivated on Friday, then deactivated again later that day, she said. On Saturday, it was reactivated a second time, only to be removed again later that day. As of Tuesday, it remains suspended.

George says that until she paid for Meta Verified — a subscription that helps creators prove their profiles are authentic — she could only interact with AI agents, which were not very helpful.

"I've spoken with quite a few other people who years ago had this happen to them, and they didn't have that [Meta Verified] option," she said. "They ended up losing their accounts because they couldn't reach somebody."

CBC has reached out to Meta for comment on this story and the company's use of AI in customer service.

AI service agents, which are significantly more advanced versions of older chatbots that could only respond to certain keywords, have become increasingly common as major AI platforms develop. 

Nevertheless, David Gerhard, who teaches computer science at the University of Manitoba, says he doubts they will ever reach the level of a human being.

"I think that they're always going to be problematic," he said.