Privacy Commissioner of Canada Philippe Dufresne announced this week that LinkedIn has agreed to voluntarily pause using the personal information of its Canadian members to train generative artificial intelligence (AI) models.
Dufresne indicated that he’s been keeping a close eye on how companies train AI at a summit earlier this year.
The development follows reporting by 404 Media in September, which found LinkedIn had begun using its users’ data for AI training without updating its terms of service. The social media platform has since updated its terms of service, and users can opt-out of their data being used. However, that won’t affect any training that’s already taken place, according to TechCrunch.
The Office of the Privacy Commissioner of Canada (OPC) said in a statement that, following the media reports, Dufresne reached out to LinkedIn to request information about the company’s training practices as well as how it obtains consent from its members.
According to the OPC, LinkedIn subsequently informed the government body that it had temporarily paused the practice while it worked to resolve the OPC’s questions. The OPC added that, while LinkedIn indicated it believed it had implemented its AI model in a privacy protective manner, it agreed to engage in discussions to ensure that its practices are compliant with Canada’s privacy laws.
“I welcome the decision by LinkedIn to pause its practice of using the personal information of Canadian LinkedIn members to train AI models while we work with them to get answers to our questions,” Dufresne said in a statement. “Personal information, even when it is publicly accessible, is subject to privacy laws and must be adequately protected.”
RELATED: Fears of repeating Web 2.0’s mistakes permeate AI-focused Competition Summit 2024
Dufresne indicated that he’s been keeping a close eye on how companies train AI at the Competition Bureau’s AI-focused summit this past September.
During a roundtable discussion of the Canadian Digital Regulators Forum at the event, Dufresne defended the parliamentary process holding up Bill C-27 and its Artificial Intelligence and Data Act (AIDA) provision, saying he hoped the bill got the priority it deserved. He went on to explain how AI doesn’t exist in a legal vacuum, and that organizations that design AI must justify their processes and uphold their current obligations under existing regulations.
“If [AIDA doesn’t pass] for whatever reason, we are acting, we have ongoing investigations, we have issued statements on AI,” Dufresne said at the time. “We’re going to continue to use the tools that we currently have.”
The federal government has been taking on big tech companies on many different fronts in recent weeks. In early November, The Government of Canada ordered social media giant TikTok to wind up its operations in Canada, citing unspecified national security risks, but did not outright ban the platform itself. TikTok challenged the order this week, alleging that federal representatives failed to substantially engage with the company during the national security review.
Canada’s Competition Bureau also sued Google last month for what it calls anti-competitive conduct in the country’s online advertising technology sector. The Bureau is seeking an order that would require Google to cease the anti-competitive practices, sell off two of its adtech tools, and pay a monetary penalty.
Feature image courtesy Office of the Privacy Commissioner of Canada via LinkedIn.