A high school history teacher in Vaughan, Ont., says she lost access to her Instagram account â and thousands of personal photos, conversations and memories â after the social media platform wrongly accused her of posting material that it said depicted “child sexual exploitation, abuse and nudity.”
Megan Conte says it took days to get through to a human at Meta, which owns Instagram, to plead her case. But even that didn’t result in her account being reinstated, she said.
“When I read what I was accused of, I was very hurt. I was very surprised, especially considering what I do for a living,” she told CBC Toronto. “And there was no one I could contact â no human.”
Conte was issued an apology by Meta, the parent company of Instagram, Facebook, WhatsApp and Threads, and her account was unlocked hours after CBC Toronto contacted the company to inquire about her complaints.
“We’re sorry we got this wrong and that you were unable to use Instagram for a while,” an email from Meta to Conte reads. “Sometimes we need to take action to keep our community safe.”
Conte said she is far from alone in her concerns about arbitrary and difficult-to-reverse decisions made by social media moderators.
An online petition started by Brittany Watson of Peterborough, Ont., complaining about a perceived over-reliance by social media companies on artificial intelligence tools rather than humans has so far garnered more than 34,000 names from people around the world.
Watson launched her campaign after she, too, was banned by Meta in May, for reasons that, she said, are still unclear. After two weeks, the ban was lifted, she said.
“Social media isn’t just social media anymore. It’s now part of daily lives,” Watson told CBC Toronto. “Now, they’re taking that away without any explanation.”
She said she’s overwhelmed by the international response to her petition. “People are just so frustrated by it.”
Watson said the point of her petition and an accompanying People over Platforms website is to force more accountability from social media sites. She’d like to see Meta rewire its moderating tools to make them more accurately detect actual wrongdoing online.
“I think the robots need to be reset,” she said.
Both Watson and Conte say they have no proof that AI is behind the mistaken bans and suspensions.
A spokesperson for Meta would not comment on how much, or how little, the company’s social media providers rely on AI to moderate members’ posts.
But London, Ont.-based tech expert Carmi Levy says it would be physically impossible for Meta to use humans alone to moderate its platforms.
“With over three billion regular users of these platforms, there’s no way that Meta could hire enough people in the world to cover everything that gets posted,” he said. “Automation is the only way they can make this scale.
“It is automation run amok.”
A Meta spokesperson told CBC Toronto the company uses a combination of people and technology to track violations of its community standards. The company also said it has not noticed an uptick in the number of people whose accounts are mistakenly suspended.
“We take action on accounts that violate our policies, and people can appeal if they think we’ve made a mistake,” the company spokesperson said.
Conte’s problems began July 26Â when she received a message from a friend pointing out that her Instagram account was inaccessible.
She logged on and discovered a note from Meta that read in part: “Your Instagram account has been suspended. This is because your account, or activity on it, doesn’t follow our community standards on child sexual exploitation, abuse and nudity.”
Even now, Conte said, she has no idea what triggered Meta’s suspension.
“The accusation is horrifying, offensive and completely false,” she told CBC Toronto. “I am a high school teacher and to be associated with such a charge … has been both traumatic and damaging.”
Making matters even more confusing, she said, is the fact she hadn’t posted anything in a couple of months.
“This goes beyond inconvenience,” she said. “I lost about 15 years of conversations, memories, business contacts, creative work and social presence. Photos of loved ones, collaborations, messages with friends â all gone in an instant because of a machine’s decision.”
All the lost content was restored when the suspension was lifted.
Conte said shutting down her account “feels like a kind of identity theft. It’s emotionally exhausting and professionally disruptive.”
After several days of trying to navigate Meta’s complaints process unsuccessfully, Conte said she eventually got through to, she believes, a human via the platform’s troubleshooting text messaging tool. But she was only able to access that feature, she said, after paying a fee to verify her mother’s Instagram account.
Conte said she had her mom’s account verified because she believes appeals on verified accounts are taken more seriously by Instagram.
However, after a text exchange that lasted more than an hour, she was still not reinstated.
Her account was only unlocked by Meta last week, after CBC Toronto’s call to the company. At the same time, she received the apology.
Levy said people who, like Conte, feel they’ve been wrongly suspended have few options.
“There’s no legislation on the books that forces the company to use humans instead of artificial intelligence for this job,” he said. “It can do as it wishes.”
