Ottawa weighs plans on AI, copyright as OpenAI fights Ontario court jurisdiction
Canada's artificial intelligence minister is keeping a close watch on court cases in Canada and the U.S. to determine next steps for Ottawa's regulatory approach to AI.
Some AI companies have claimed early wins south of the border, and OpenAI is now fighting the jurisdiction of an Ontario court to hear a lawsuit by news publishers.
Evan Solomon's office said in a statement he plans to address copyright "within Canada's broader AI regulatory approach, with a focus on protecting cultural sovereignty and how [creators] factor into this conversation."
But there are no current plans for a standalone copyright bill, as Solomon's office is "closely monitoring the ongoing court cases and market developments" to help chart the path forward.
It's unclear how long it will take for those court cases to determine whether artificial intelligence companies can use copyrighted content to train their AI products.
The sole Canadian case to pose the question was launched late last year by a coalition of news publishers and the Ontario Superior Court is set to hear a jurisdictional challenge in September.
The coalition — which includes The Canadian Press, Torstar, the Globe and Mail, Postmedia and CBC/Radio-Canada — is suing OpenAI for using news content to train its generative artificial intelligence system.
The news publishers argue OpenAI is breaching copyright by scraping large amounts of content from Canadian media, and then profiting from the use of that content without permission or compensation.
They said in court filings that OpenAI has "engaged in ongoing, deliberate and unauthorized misappropriation of [their] valuable news media works."
"Rather than seek to obtain the information legally, OpenAI has elected to brazenly misappropriate the News Media Companies' valuable intellectual property and convert it for its own uses, including commercial uses, without consent or consideration."
OpenAI challenging jurisdictionOpenAI has denied the allegations, and previously said its models are trained on publicly available data, and "grounded in fair use and related international copyright principles."
The company, which is headquartered in San Francisco, is challenging the jurisdiction of the Ontario court to hear the case.
It argued in a court filing that it's not located in Ontario and does not do business in the province.
OpenAI also argued the Copyright Act doesn't apply outside of Canada.
OpenAI is asking the court to seal some documents in the case. The court is scheduled to hold a hearing on the sealing motion on July 30, according to a schedule outlined in court documents.
It asked the court to seal documents containing "commercially sensitive" information, including about its corporate organization and structure, its web crawling and fetching processes and systems, and its "model training and inference processes, systems, resource allocations and/or cost structures."
"The artificial intelligence industry is highly competitive and developing at a rapid pace," says an affidavit submitted by the company. "Competitors in this industry are many and range from large, established technology companies such as Google and Amazon, to smaller startups seeking to establish a foothold in the industry.
"As recognized leaders in the artificial intelligence industry, competitors and potential competitors to the defendants would benefit from having access to confidential information of the defendants."
A lawyer for the news publishers provided information on the court deadlines, but did not provide comment on the case.
Numerous lawsuits dealing with AI systems and copyright are underway in the United States, some dating back to 2023. In late June, AI companies won victories in two of those cases.
In a case launched by a group of authors, including comedian Sarah Silverman, a judge ruled AI systems' use of published work was fair use and the authors didn't demonstrate that use would result in market dilution.
But the judge also said his ruling affects only those specific authors — whose lawyers didn't make the right arguments — and does not mean Meta's use of copyrighted material to train its systems was legal. Judge Vince Chhabria noted in his summary judgment that in "the grand scheme of things, the consequences of this ruling are limited."
In a separate U.S. case, a judge ruled the use by AI company Anthropic of published books without permission to train its systems was fair use. But Judge William Alsup also ruled Anthropic "had no entitlement to use pirated copies."
Jane Ginsburg, a professor at Columbia University's law school who studies intellectual property and technology, said it would be too simplistic to just look at the cases as complete wins for the AI companies.
"I think both the question of how much weight to give the pirate nature of the sources, and the question of market dilution, are going to be big issues in other cases."
cbc.ca
