Canada’s minister responsible for artificial intelligence revealed that the CEO of OpenAI has agreed to allow Canadian experts access to the company’s safety office to assist in evaluating potential threats following the tragic incident in Tumbler Ridge, British Columbia.
During a virtual meeting, Evan Solomon discussed with OpenAI’s leader, Sam Altman, for half an hour on Wednesday. OpenAI is the owner of ChatGPT, a popular artificial intelligence chatbot.
Solomon shared that Altman expressed horror and a sense of responsibility for failing to flag concerning content. This led to the company initiating changes to prevent such oversights. Solomon’s insights were shared in an interview with CBC’s Power & Politics.
The focus of the discussion was on the ChatGPT account associated with the shooter, which OpenAI had banned and internally flagged eight months before the tragic incident. Despite the account’s posts related to gun violence, OpenAI did not inform authorities until after the events, citing that the activities did not meet their reporting criteria at the time.
B.C. Premier David Eby has demanded an apology from the CEO to the community of Tumbler Ridge and is scheduled to meet with Altman on Thursday.
Solomon conveyed his support for Premier Eby’s call for an apology, emphasizing the government’s backing of the request. Altman has not yet made any public statements regarding his company’s involvement in the recent events.
In a late-night statement, an OpenAI spokesperson mentioned that Altman discussed with Solomon the measures being taken by the company, including enhancing law enforcement referral criteria and adapting systems to consider regional contexts more effectively.
The minister also disclosed that Altman agreed to involve Canadian experts in mental health and law within OpenAI’s safety office, and to allow the Canadian AI Safety Institute to conduct a comprehensive evaluation of the new safety protocols. Additionally, OpenAI committed to providing a detailed report on the enhanced systems for identifying high-risk individuals.
Solomon urged OpenAI to report threats directly to the Royal Canadian Mounted Police (RCMP) in addition to the FBI. He stressed the importance of stringent safety protocols to ensure the well-being of Canadians and emphasized the need for transparent regulatory frameworks.
Amid calls for AI regulation following the incident, Solomon affirmed that all options are being considered by the federal government. He mentioned plans to engage with other platforms to verify the implementation of adequate safety measures, without specifying the companies involved or confirming any discussions that may have taken place.
The recent discussions follow a previous meeting where Solomon expressed disappointment with OpenAI’s commitments to enhance safety measures. Despite assurances from the company, Solomon noted a lack of clarity on the proposed changes and their implementation plan.
In a troubling revelation, OpenAI disclosed that the shooter, Jesse Van Rootselaar, had created a second ChatGPT account after being banned. The tragic incident involved Van Rootselaar taking the lives of her family members and several others before the heartbreaking end at the local school.
The ongoing dialogue and commitments from OpenAI signal a collaborative effort with the Canadian government to address safety concerns and ensure better protection for the public.