OTTAWA – Canada’s Artificial Intelligence (AI) Safety Institute has gained access to all of OpenAI’s protocols, according to Artificial Intelligence Minister Evan Solomon, who made this announcement on Friday. This development comes as the institute works on a report aimed at ensuring accountability in the AI sector.
Minister Solomon's remarks followed serious concerns that arose after OpenAI banned a mass shooter from using its ChatGPT chatbot but failed to alert law enforcement about the individual’s dangerous behavior. This individual, Jesse Van Rootselaar, managed to circumvent the ban by creating a second account, resulting in a tragic incident in Tumbler Ridge, British Columbia, on February 10, where she fatally shot eight people, including six children, before taking her own life.
During a meeting with OpenAI’s CEO in March, Solomon expressed the necessity for heightened responsibility in AI safety measures. He emphasized that the government is determined to protect children and warned that “the hammer comes down” if such protections are not upheld. The conversation underscored the urgent need for AI companies to prioritize safety and to put systems in place that would prevent similar tragedies in the future.
In the aftermath of the meeting, Solomon instructed the Canadian AI Safety Institute to assess OpenAI's operational model and to provide his office with expert technical advice. OpenAI representatives assured him that they would implement a range of new measures, including the creation of a report that outlines systems aimed at identifying high-risk offenders and policy violators. They have also committed to establishing a direct line of communication with the Royal Canadian Mounted Police (RCMP) and developing safety protocols that would guide individuals experiencing distress to local assistance services.
Minister Solomon made these statements while discussing the government’s proposed online harms bill at the Liberal party policy convention in Montreal on Friday. However, he did not clarify whether this legislation would extend to cover AI chatbots specifically, leaving open questions about the scope of regulatory measures aimed at mitigating risks associated with AI technologies.
This situation highlights the increasing scrutiny that AI companies face regarding their safety protocols and responsiveness to potential threats. The discussions and planned improvements outlined by both the Canadian government and OpenAI reflect a growing recognition of the responsibility that technology companies bear in ensuring user safety and ethical interaction. As AI technologies continue to evolve, regulatory authorities worldwide are likely to implement more stringent measures to protect vulnerable populations, particularly children, from potential harms associated with emerging technologies.
