With much still left to be determined in the federal government’s newly proposed social media restrictions, experts and officials are parsing the details to figure out how the legislation might work, where the roadblocks could be, and what’s missing.
B.C. Attorney General Niki Sharma says the province wants more, including rules that her government had proposed forcing artificial intelligence companies to report activity to police when someone is threatening themselves or the community.
“It does fall short of what we asked for,” she said.
The Safe Social Media bill, introduced on Wednesday (June 10), would restrict social media companies from allowing users under 16 to have an account or access inappropriate material, as well as give the government new takedown power to force the removal of harmful posts. It would also create a digital safety commissioner to oversee the regulatory structure.
The bill also leaves many of the details to be worked out later.
Sharma wants to be at the table for these discussions, particularly after the Tumbler Ridge shootings. The suspect’s OpenAI account was flagged by the company last summer, but nothing was reported to police until after the February shooting.
“What we want to make sure is that the rules that are in place and the regulatory oversight that’s in place would be strong enough to prevent some of the tragedies that we’ve seen,” she said.
The bill’s vagueness on implementation could help regulators contend with the changing spectrum of issues involved.
Chris Tenove, the assistant director of the Centre for the Study of Democratic Institutions at the University of British Columbia and an expert on social media policy, suggests this may be necessary as many of the tools needed to implement these restrictions are in development.
Tenove called the social media age restrictions and age-related safeguards for AI chatbots in particular “ambitious,” with many of the key details still to be worked out.
In terms of age verification, he said there are several options on the table, from requiring government identification from everyone to using third-party age-verification companies to utilizing age-estimation video software.
“What will actually be the standard remains to be seen,” Tenove said.
There could also be technical limitations that result in pushback from platforms, or concerns raised during trade negotiations with the United States, where many of these companies are based.
These seem to be at least some of the reasons for the flexibility baked into the bill.
“It’s like a blueprint that gives the sense of the house to be built, but there’s uh a lot that can be done to accommodate what neighbours and future residents might want from them,” Tenove said.
Ultimately, the bill also avoids seeking “perfection” from companies, but rather to add enough “friction” to reduce use by those under 16.
“This bill is, to a significant extent, a piece of leverage to really push the technology companies to have stronger child safety guidelines,” Tenove said.
On the AI chatbot side, he said the bill seems to focus on reducing harms associated with users who form relationships with chatbots. He called this a “pretty novel” proposal, with few examples in other jurisdictions of specific regulation governing chatbots that “anthropomorphize” and pretend to be human-like. Defining what’s included here could be difficult.
“It’s ambitious, and it could apply to a lot, but they’ve also tried to scope it down to certain types of really pernicious behaviour reminiscent of what we’ve seen in Tumbler Ridge,” he said.
Another interesting bit of the bill, he said, is the obligation for social media companies to add a “synthetic” label to AI-generated content.
“Again, there’ll be lots of mistakes and miscases when people disguise stuff,” Tenove said. “But it’s raising the bar.”
The bill still stops short of forcing companies to report online activity to police, possibly because of the tricky questions involved. Tenove said there is a “tension” on this question between ensuring police are warned of impending danger and protecting privacy, as well as preventing a flood of reports police must then follow up on.
“It can sound too easy to say if there’s really concerning material, you just send the police a note,” he said.
B.C. Premier David Eby isn’t satisfied, calling this reporting requirement a “key recommendation” that came out of the Tubler Ridge tragedy. He plans to keep pushing for a minimum reporting threshold.
“It is essential for the safety of Canadians, and we can’t leave these companies to regulate themselves on this crucial piece that has already visited such harm on British Columbians,” he said.