Canada Moves to Ban Social Media for Minors and Tighten AI Oversight

Canada may soon become another country to sharply restrict children and teenagers’ use of social media: the government introduced the “Safe Social Media Act” this week, a new bill that would bar people under 16 from opening social media accounts and impose a series of new obligations on digital platforms and artificial intelligence companies. If approved, Canada would join countries such as Australia, Indonesia and Malaysia, which have already advanced similar measures aimed at reducing the impact of social media on children. The Canadian bill is part of a broader wave of similar initiatives around the world, though it is still too early to assess their actual effect.

According to the proposal, companies would be required to design their services so they are safer for minors and to quickly remove content containing AI fakes, or material depicting child sexual exploitation or harming victims of sexual offenses. In addition, the law is expected to require clear labeling of content created with artificial intelligence, allow simple reporting of harmful and violent content on social networks, and give users more convenient tools for blocking problematic accounts. At the same time, the bill also includes what is described as a first-of-its-kind reference to AI-based chatbots. Unlike social networks, Canada does not intend to restrict access to AI systems by age, but rather to require AI platforms to take steps to prevent the spread of harmful content through chatbots, reduce behavior that could cause harm to users, and operate emergency mechanisms for crisis situations. Canadian Culture Minister Marc Miller explained that “the harm caused by social media has been studied for many years, while the effects of chatbots are not yet understood to the same degree, and they also do not play the same social role. Therefore, we saw no reason at this stage to restrict access to them.”

According to foreign reports, these clauses were added following one of the deadliest shooting incidents Canada has seen in recent years, which took place last February at a school in the town of Tumbler Ridge in British Columbia. After the incident, it emerged that the shooter had held conversations with ChatGPT months earlier that included violent scenarios and the use of weapons. The conversations were flagged by OpenAI’s safety systems and also sparked internal discussion among more than a dozen employees about whether to alert law enforcement. In the end, the company decided to block the account but not report it to police, saying it had not identified an “imminent and credible threat.” After the incident, the company faced harsh criticism, and a lawsuit was filed against it by the family of one of the victims, who argued that the tragedy could have been prevented if it had issued a warning in time. This is essentially the main reason the new Canadian bill includes explicit obligations on AI platforms for the first time, and not only on social networks.