Canada may regulate AI chatbots if OpenAI fails to improve safety protocols, following scrutiny over unreported posts by a mass shooter
OTTAWA: The Canadian government has warned it will consider new regulations for artificial intelligence chatbots if OpenAI does not swiftly enhance its security measures.
This follows intense scrutiny over the company’s failure to alert police about online posts made by a mass shooter months before an attack.
Senior OpenAI officials were summoned to Ottawa to explain why they did not notify the Royal Canadian Mounted Police about activity from Jesse Van Rootselaar’s account.
The 18-year-old transgender woman killed eight people in Tumbler Ridge, British Columbia on February 10.
OpenAI confirmed it shut down her account in June due to posts related to violent activity but said nothing indicated an imminent attack.
Justice Minister Sean Fraser delivered a firm message during the meeting with company executives.
“The message that we delivered, in no uncertain terms, was that we have an expectation that there are going to be changes implemented, and if they’re not forthcoming very quickly, the government’s going to be making changes,” Fraser told reporters.
In a statement, OpenAI called the Tumbler Ridge shootings “an unspeakable tragedy.”
The company said it has strengthened safeguards and changed its law enforcement referral protocol for violent activities.
“But the ministers underscored that Canadians expect continued concrete action and we heard that message loud and clear,” the spokesperson’s statement added.
OpenAI committed to providing an update on additional steps in the coming days.
Van Rootselaar killed eight people in the remote mining town, which has a population of 2,400.
She first killed her mother and brother at the family home before heading to the local secondary school.
There, she shot dead five children and a teacher before dying from a self-inflicted gunshot wound.
Canada’s Minister for Artificial Intelligence, Evan Solomon, summoned OpenAI after learning it had banned the shooter’s account months prior.
He described the decision not to inform law enforcement as “very disturbing.”
OpenAI stated it maintains a very high threshold before deciding to involve police about alarming online posts.
