When Jasmine Ault woke in the middle of the night last Thursday and checked her phone, she found a notification from Meta saying her Instagram and Facebook accounts had been permanently disabled, and she has not reached a human being at the company since.
The deletion erased 20 years of personal photographs and professional contacts. Ms. Ault, who lives in Toronto, opened her Facebook account in 2006, the first year it was public. Gone, too, was a separate Instagram account she ran for her dog. She stayed up the next 26 hours straight, trying everything she could to appeal.
“It feels like I’ve lost everything in a metaphorical fire with my digital albums and my digital memories,” she told CBC News. “And who’s going to protect those if I can only speak to AI?”
Ms. Ault is the latest case of a pattern that has grown more common in Canada over the past two years, in which Facebook and Instagram users wake up to find their accounts gone and their only recourse a chatbot. One Canadian lost 18 years of data after posting photos from a family trip, and a second account was banned when she tried to return. Others lost accounts they depended on to advertise their businesses. Many believe Meta’s automated moderation flagged them in error; some regained access after going to reporters, while others never did.
Screenshots Ms. Ault shared with CBC News said her account “doesn’t follow our rules.” She said she was never told which rules. She believes her Facebook had recently been hacked, which may have set off the flag. Within minutes of pressing the appeal button, she learned her case was closed and she could not appeal again. She had no chance to make her argument.
So she made a backup Instagram account and paid for Meta Verification, believing the enhanced support that came with the Standard tier, $14.99 American, or about $21 Canadian, would connect her with a live agent. What she got was another chatbot, which closed her ticket and told her she had used up her support interactions. Only the Premium and Max tiers, at $150 and $500 American a month, even mention the ability to “request a call from an agent.”
“It, to me, feels like Meta is trying to cut corners by not paying a human, when they’re worth billions and billions of dollars,” Ms. Ault said. “And it feels almost sick to me that they just think, like, ‘We can just shut these people’s accounts down, they can’t get through to us and talk to us,’ and it’s people’s lifelines. It’s people’s memories. It’s people’s everything.” She said she was losing work for her side business, which markets to clients through social media.
Meta did not respond to questions from CBC News about Ms. Ault’s account, or about whether talking to a human — once the starting point of any customer service interaction — now requires a premium subscription. The AI Minister, Evan Solomon, has not answered questions about whether coming legislation will guarantee Canadians the right to reach a person during an appeal.
The frustration runs wider than Meta. A survey released this year of more than 34,000 customers, service representatives and executives worldwide, commissioned by the software company ServiceNow with the research firm ThoughtLab, found that half of customers called a lack of empathy their biggest frustration with AI customer service. Where 87 percent of customers ranked phone calls a preferred channel, only 7 percent of executives planned to prioritize them. The survey included 1,350 customers and 195 executives in Canada.
Ela Veresiu, an associate professor of marketing at York University’s Schulich School of Business, said companies adopt AI service despite the backlash because it is cheaper than staffing phones around the clock. Chatbots can recite a business’s hours, she said, but cannot answer a complex “why” question. And while customers of an ordinary company can walk away from bad service, she said, Meta’s users have nowhere to go.
Jon Festinger, a lecturer at the Peter A. Allard School of Law at the University of British Columbia, said users who relied on their accounts for business may have a claim in court for reputational damage. Canadian courts have touched the subject once before, holding Air Canada liable in 2024 for advice its chatbot invented. But gaps in consumer law remain.
Shion Guha, an assistant professor at the University of Toronto and director of its Human-Centered Data Science Lab, said the real problem is not that automated moderation errs but that no one corrects the error. “Literally in every other facet of our life, if we are appealing ... a potentially unfair decision, we reach a human being,” he said, citing speeding tickets and tax appeals. He pointed to the European Union’s Digital Services Act, in force since 2024, which guarantees users a full appeal when content is removed.
New Canadian legislation on AI chatbots, introduced in June, covers safety rules aimed at protecting children online. It says nothing about consumer rights.