Toronto girl screams Meta deleted account. No caregiver, only AI.
Toronto girl screams Meta, erases overnight account, can't contact authorities, relies on AI, even pays for service
Several Canadians complained were shut down by Meta for unknown reasons, while an AI-based appeal and customer service system prevented access to human officials; experts called for consumer protection rules and fair channels of appeal.
Jasmine Alt, a native of Toronto, Canada, faced major anxiety after waking up in the middle of the night to find that her Facebook and Instagram accounts had been permanently disabled by Meta, with no clear explanation of which rules she had violated.
Alt has been using Facebook since 2006 and has used social media both to store images of personal memories, contact people, as well as to market them to her ancillary business. She stated that information and contacts accumulated over the years are highly important to life and work.
The latter discovers that her account has been closed, Alt tries to appeal, but instead finds that the system acted quickly, without giving her the opportunity to elaborate on the situation, before informing that the case has been closed and no further appeal can be filed.
Alt also found that before the account was closed, several unsuccessful attempts to log into her Facebook had been made, leading her to suspect that the account might have been hacked and that the incident may have been related to Meta's detection system misidentifying her account.
Pay Meta Verified, but also talk to AI.
To find a way to contact the authorities, Alt created a new Instagram account and signed up for the standard Meta Verified service, which costs US $14.99 per month, understanding that it will receive better support.
But what she got back was a conversation with an AI chatbot like before.
The system says her petition will be closed and she has exercised her right to seek a specified amount of assistance, leaving her still unable to contact human authorities to explain the problem.
Alt said she felt like she had lost both her memory and the business network that had accumulated over the years, especially her photos, conversations and contacts on the platform.
For higher levels of Meta Verified services, a callback benefit is specified, but up to US $150 per month is required for Premium and US $500 for Max levels.
Meta did not answer CBC News' questions about Alt's case, including the question of how direct human officer contact is currently limited.
Not a single case. Many Canadians face the same problem.
Alt's case came amid complaints from a number of Canadian users over the past two years, who stated that their Facebook or Instagram accounts had been abruptly closed.
Some have lost information accumulated over many years, while others have lost accounts used for advertising and business.
Some users believe that Meta's automated monitoring system may have mistakenly detected accounts, with some being able to recover accounts after receiving media attention, while others still have no access to their own accounts.
The issue reflects broader questions about the adoption of AI as a human substitute in customer service work, particularly in cases involving key decisions and appeals.
Customers want humans, but companies increase the use of AI.
A global survey conducted by ServiceNow and ThoughtLab, which surveyed over 34,000 customers, service representatives and executives, found that 50% of customers identified lack of empathy as the most offensive to AI-based customer service.
Meanwhile, 87% of customers say that telephone contact is the most desirable service channel, but only 7% of executives say that their company plans to focus on it.
Ela Vereziu, an associate professor of marketing at the Schulich School of Business, York University, sees that many companies continue to accelerate the adoption of AI because it can reduce costs and increase efficiency.
However, AI may suit simple questions such as business opening hours, but when faced with complex problems that need to explain "why" the decision was made, automation may not be able to meet the needs of users.
For large platforms with Meta sample-specific services, users also have limited options to move to other services, which makes the business pressure to change the customer service system potentially different from businesses with many direct competitors.
Experts point out. Users should have the right to appeal to humans.
According to Jon Festinger, a professor at the University of British Columbia's Peter A. Allard School of Law at the University of British Columbia, Canadians may have legal avenues to object to account deletion and claim damages, especially in cases where accounts are used for business purposes.
He gave the example of Air Canada's 2024 case, in which a Canadian court considered issues regarding the information that chatbot systems provide to customers; and saw Canadian law as still having gaps in consumer rights as AI technology came to play a greater role.
Shion Guha, an assistant professor at the University of Toronto and director of the human-centered Data Science Laboratory, said that all types of technological systems can make mistakes, so it's important not just to make the system go wrong less, but to have a mechanism for the affected to be able to correct mistakes.
He was of the view that companies should have clear channels for users to contact human authorities and that there should be a fair appeals process, especially when decisions are made by automation.
The EU has a legal framework, while Canada is considering new legislation.
One of the examples that has been compared is the Digital Services Act, which applies to a broad range of online platforms since 2024 and establishes transparency requirements, including mechanisms to dispute content decisions.
Canada, meanwhile, is considering a bill related to AI and decision-making by automation.
According to Minister Evan Solomon's office, Bill C-36, proposed in June, would require individuals to be able to seek explanations of decisions by automation that could have significant legal or impact effects, and to be able to clarify the facts to officials with the authority to review such decisions.
However, the bill has not yet been approved, nor does it directly stipulate that the company must provide a human customer service contact channel.
For Alt, what she needs now is not just an explanation, but an opportunity to talk to real people to clarify what happened to her account.
Because for many users, Facebook and Instagram are no longer just social media platforms, but a collection of images, memories, business networks, and life records that have accumulated over the years. And when automation decides to close accounts, the key question is not only whether AI makes the right or wrong decisions, but when AI goes wrong, who can users claim and talk to?
จากประสบการณ์ของผู้ใช้หลายคนในยุคที่เทคโนโลยี AI มีบทบาทในงานบริการลูกค้า บางครั้งการพึ่งพาระบบอัตโนมัติอาจนำมาซึ่งปัญหาเมื่อต้องจัดการกับกรณีที่ซับซ้อน เช่น การถูกปิดบัญชีโดยไม่มีคำอธิบายเหมือนในกรณีของสาวโตรอนโตที่ถูก Meta ลบบัญชีแบบไม่แจ้งล่วงหน้า ผู้ใช้บางคนเจอปัญหาเข้าไม่ถึงเจ้าหน้าที่จริงๆ ทำให้ปัญหาของพวกเขาไม่ได้รับการแก้ไขอย่างตรงจุด การสนทนากับ AI ไม่ได้ให้ความรู้สึกว่ามีใครรับฟังหรือเข้าใจความกังวลของผู้ใช้จริง โดยเฉพาะเมื่อข้อมูลส่วนตัวและธุรกิจที่สะสมมาเป็นเวลานานขึ้นอยู่กับบัญชีเหล่านั้น ทั้งนี้ การตัดสินใจของ AI โดยไม่มีมนุษย์เข้ามาทบทวน หรือช่องทางอุทธรณ์ที่เป็นธรรม อาจทำให้ผู้ใช้รู้สึกหมดหวังและสูญเสียทั้งความทรงจำและรายได้ ผู้ใช้จึงต้องการช่องทางที่ชัดเจนสำหรับติดต่อเจ้าหน้าที่มนุษย์เพื่ออธิบายเหตุผลและมีกระบวนการแก้ไขข้อผิดพลาดที่เหมาะสม ในยุโรปมีกฎหมายที่บังคับใช้กับแพลตฟอร์มออนไลน์เพื่อให้มีความโปร่งใสและกลไกโต้แย้ง แต่หลายประเทศยังไม่มีข้อบังคับดังกล่าวอย่างเต็มที่ ซึ่งประเทศแคนาดากำลังพิจารณากฎหมายที่เกี่ยวกับ AI เพื่อคุ้มครองสิทธิผู้บริโภคในประเด็นนี้ จากเรื่องนี้จึงเป็นเสมือนคำเตือนให้บริษัทเทคโนโลยีขนาดใหญ่ที่ใช้ AI ในบริการลูกค้า ควรเพิ่มทางเลือกให้ผู้ใช้ในการติดต่อกับเจ้าหน้าที่จริงๆ และทำให้กระบวนการอุทธรณ์มีความเป็นธรรม เพื่อป้องกันการสูญเสียที่ไม่ควรเกิดขึ้นของข้อมูลและความไว้วางใจจากผู้ใช้บริการ
