While some praise Manitoba’s promise to ban youth-driven artificial intelligence chatbots and social media, others point out how difficult it will be to implement effectively.
Premier Wab Kinew announced over the weekend that Manitoba will become the first province in Canada to ban youth under 16 from using social media platforms such as Instagram and TikTok. It also announced it would go further than jurisdictions such as Australia by banning the use of AI chatbots such as ChatGPT and Claude by young people.
“The way we’ve approached this problem so far has been to replace one risk with an even bigger risk,” said Carmi Levy, a technology analyst and journalist based in London, Ont.
“By including AI on the list of technologies subject to this ban, states are recognizing that, yes, this is a new and emerging threat. Our laws need to cast a much wider net.”
Some jurisdictions considering bans are primarily focused on what emerging science suggests about the harm to the developing brain associated with social media use.
It would also be worth putting limits on access to chatbots and creating reporting requirements for big tech companies to flag problematic uses of their technology, said Emma Durden, Canada Research Chair in Neuroscience and Learning Disabilities.
While some have praised Manitoba for becoming the first Canadian jurisdiction to take steps to block social media and AI chatbots, experts question whether it will be effective.
Duerden said Manitoba’s decision to ban AI chatbots aimed at young people could be a “world first.”
She pointed to a mass murder in Tumbler Ridge, British Columbia, involving a teenage gunman who had been banned from ChatGPT months before the massacre due to questions involving gun violence.
OpenAI, the company behind ChatGPT, “could have issued a warning and could have alerted authorities in advance,” said Durden, an associate professor at Western University’s School of Education.
“I think this really reveals what Canada is capable of right now.”
Duerden said her research and that of others suggest that social media use is associated with mental health risks in the developing brain. He said children are four to seven times more likely than adults to become addicted to certain behaviors or drugs.
Duerden’s team studied 500 teens in Canada and the United States and found that passive scrolling was associated with increased levels of clinical anxiety compared to creating content and connecting with others. The study shows the benefits of limiting passive scrolling to two hours a day, she said.
”[These platforms] They are not neutral…they can affect the developing brain, attention, learning and memory, and mental health,” Duerden said.
“You can’t hold parents responsible. You can’t hold teachers responsible. You can’t hold individual children responsible. What we need at this point is actual government legislation to limit access.”
Azfar Adib, a computer science expert, specializes in age verification technology. He would like to see a Canada-wide implementation strategy, but said, “Unfortunately, there is no legislation in place at the federal level in Canada.” [has] It’s going pretty slowly. ”
Adib said the first widely used age verification methods elsewhere included having users upload government identification such as a driver’s license or health card.
LinkedIn is one platform that asks users to take selfies, which the AI scans to estimate their age range, he said.
New technology is also emerging that uses audio recordings and hand and arm movements to determine age.
Since heart rate changes with age, Adib is working on an innovation that leverages electrocardiogram technology to measure a user’s pulse and determine their age based on that.
“New solutions are emerging in what we call ‘privacy by design,'” he said.
“When we talk about age verification around the world, the biggest concern is privacy, because not everyone wants to share their ID or even a selfie.”
Legal concerns are also involved.
“This fundamentally changes the nature of the surveillance infrastructure that will exist in the future, expands it significantly, and opens up all sorts of new vectors for attacks and social engineering,” said Michael Karanikolas, associate professor of law at Dalhousie University in Halifax.
“If there’s one thing I can advise the government… it’s to talk to the information security community, to the cybersecurity experts, to make sure that we have a good and robust understanding of the impact of these technologies on digital privacy and digital surveillance, both among young people and the population as a whole.”
Karanikolas points out that while systems that require some form of government ID are not foolproof, such systems are more effective than the honor systems in place in some apps.
“But of course, if you’re … letting these platforms and AI companies collect the passports and identities of all your users, that’s going to be a significant amount of highly sensitive personally identifying information that these platforms are going to collect,” he said.
Levy, a technology analyst and journalist, said some children will find ways to circumvent the verification, but that won’t stop the government from trying.
“At the very least, this puts us on the roadmap towards enacting laws that protect them from the worst kinds of abuse,” he said. “Anything is better than nothing.”
