Open this photo in gallery:
Canadian Identity Minister Marc Miller takes his seat for a news conference in Ottawa on Wednesday.Adrian Wyld/The Canadian Press
Social-media companies will be required to label content created using artificial intelligence, as well as harmful posts generated by bot farms, under Ottawa’s Safe Social Media bill introduced Wednesday.
The move follows the proliferation of computer-generated images and audio online that range from videos of tigers making friends with kittens to realistic AI-generated deepfakes used by Russia to spread disinformation about Ukraine.
The platforms will also have to label harmful content, including that fomenting hatred or extremism, or inciting violence, that are pushed out by automated bot farms. Some of these are sponsored by foreign states with the intention of spreading disinformation, sowing dissent and distrust in governments.
Introduced Wednesday by Canadian Identity Minister Marc Miller, the bill also includes a social-media ban for users below the age of 16, with exemptions for companies that meet standards set by the Digital Safety Commission, a new regulator to be established by the bill.
Ottawa introduces bill to restrict social media for teens, regulate AI chatbots
The bill acknowledges that identifying AI content can be difficult, even though AI can itself be harnessed to detect synthetic images, with mixed results.
Bill C-34 says the Digital Safety Commission must consider the extent to which it is “technically feasible” for social-media companies to identify synthetic content before it can determine whether their efforts to do so are adequate. The criteria for labelling synthetic content will be determined by the government through regulations.
“Harmful synthetic content, including sexually explicit deepfakes, can have devastating and lasting impacts on victims and their families,” said Alisson Lévesque, spokesperson for Mr. Miller.
She said the bill will “require social media platforms to put in place robust safeguards to protect children on their services, including labelling bot-driven harmful content and all synthetic content, and reducing children’s exposure to pornographic content.”
Ben Waldman, chief executive of Gander Social, a new Canadian platform, said: “We already ask people on Gander to label AI-generated content, and we support that kind of transparency when it helps people understand what they’re seeing.”
One of the roles of the new commission will be to determine whether platforms covered by the act are complying. Companies that breach the act or flout an order from the commission may have to pay a maximum penalty of either $10-million or 3 per cent of gross global revenue, whichever is larger.
Under the bill, larger social-media platforms – likely including X, TikTok and Facebook – will have to block users younger than 16 from having accounts. But the platforms would be able to opt out of the under-16 ban if they satisfy the commission that they have introduced sufficient safeguards to protect children from harm.
The bill follows the introduction six months ago of a social-media ban in Australia for users under 16, which some experts have warned is being flouted by a large number of teens, including some using smaller platforms not covered by the ban.
Opinion: Ban social media for young people? Good luck with that
Canada’s bill requires that the Canadian Identity Minister review the entire act on its third anniversary to see how it is operating, and specifically how the social-media ban for children younger than 16 and the exemption system for companies are working.
Emily Laidlaw, Canada Research Chair in cybersecurity law at the University of Calgary, said requiring a review after three years was “a good thing.”
“I think this is important because it shows they want to see if this is working or not, and it also softens the impact of those provisions,” said Prof. Laidlaw, who served on an expert committee that advised Mr. Miller on the possible content of the bill. “It’s not like they are saying, ‘we came up with the solution,’ and walking away. They are acknowledging it’s an iterative process.”
The bill would also require social-media platforms that carry pornographic content to verify that their users are at least 18.
This would likely include sites such as Pornhub that feature uploaded adult videos. Precisely which platforms carrying adult content will be covered by the bill will be defined through regulations.
The bill says that both age verification, which could require presenting government identification, and age estimation measures, which can include gauging a person’s age by scanning their face, hand or arm, could be used. However, the measures must be effective, the bill specifies, with steps to protect privacy.
The Liberal government’s new online safety legislation would force social media platforms to block access for kids under 16 and would regulate chatbots.
The Canadian Press
Britain’s Online Safety Act, which became law in 2023, requires porn sites to verify that their users are 18 or older. Ofcom, the British communications regulator, sets out criteria and standards for age verification, but it does not specify the precise method that companies should use.
Pornhub, which is owned by Canadian private equity firm Ethical Capital Partners and operated by Aylo, withdrew from the British market in January over the age-verification requirement.
But last month, the platform reintroduced access in Britain for users of the latest version of Apple’s iOS, which allows adults to verify their age through their Apple account.
Alex Kekesi, head of community and brand at Aylo, said in an e-mail: “The recent UK Apple iOS update is a huge step in the right direction toward age verification that actually works. Aylo is glad to see this roll out in the UK and looks forward to more governments mandating effective child protection measures like this.”
Michael Geist, the University of Ottawa’s Canada Research Chair in internet and e-commerce law, said some social-media platforms might have to verify their users’ ages twice – once to determine whether they are 16 or older and, if they feature pornographic content, a second time to verify that they are at least 18.
