The Minister of Innovation, Pierre Fitzgibbon, wants to oversee the development of artificial intelligence (AI) in order to avoid the slippages feared by several experts.
Mr. Fitzgibbon has invited experts and opposition representatives to a meeting to discuss the issue next week in Montreal.
“The regulatory framework, the ethical framework must be better defined, he declared Thursday during the question period. We are lucky to have people in Quebec who will be able to help us. »
Last week, Quebec researcher Yoshua Bengio, who will participate in the meeting, co-signed a letter in which a six-month moratorium is requested from artificial intelligence laboratories. Governments were also invited to work on an AI governance system.
“We are very concerned about the societal issues that have been described in the letter, particularly when machines replace humans,” Fitzgibbon said Thursday. Clearly, we want to have a non-partisan exercise so that all participants can contribute. »
The federal government had already signaled its intention to better regulate AI development activities last week.
bad endings
Six AI experts in Quebec, including Mr. Bengio, will offer their perspective on Wednesday to allow Quebec to have “a dominant position in the international market”, Mr. Fitzgibbon said Thursday.
“We are going to start working with people from the industry,” he said.
After the question period, Mr. Fitzgibbon said that he will make an announcement on this occasion, the content of which he did not reveal.
“We don’t want our researchers to do things in terms of algorithms that will be used for bad purposes, we have to control the use of AI,” he told reporters. In Quebec, we have no issues on that right now. »
His colleague Sonia LeBel, president of the Treasury Board, was not informed of problems with artificial intelligence in the public service.
“I haven’t been alerted to any misuse of anything,” she said. We are very concerned about the ethical framework. »
During question period, Liberal MP Fred Beauchemin, who will take part in the meeting convened by Mr. Fitzgibbon, said that a parliamentary committee must take up the matter.
“There are known limitations about artificial intelligence like ChatGPT, biases like racism, sexism, the use of such software for misinformation, data privacy, impact on studies and the market labor, legal liability and copyrights,” he said, after asking a first question entirely prepared by ChatGPT.
The Minister of Cybersecurity and Digital, Eric Caire, said that there are already provisions on the subject. “We also adopted the first artificial intelligence integration strategy which considers the ethical aspects of artificial intelligence,” he said in the House.