In the space of a few hot September days, Canada’s Minister of Innovation, Science and Economic Development, François-Philippe Champagne, went from food specialist to artificial intelligence expert. His angle: I represent 40 million Canadians, you have to listen to me.
The energetic federal minister met with the leaders of large Canadian companies to kindly ask them to be more respectful of Canadian consumers. The major grocers immediately promised to act in the direction requested by François-Philippe Champagne. Price freezes, more discounts, etc.
Nothing, basically, that we hadn’t already seen before.
A few days later, the minister published a voluntary Code of Conduct on the adoption, use and development of artificial intelligence (AI) applications. This code includes six main principles, expressed in a fairly general way, but, as it is voluntary, it is not accompanied by any constraint, any verification or any penalty if ever a company which claims to comply with it conform commits an oddity.
Minister Champagne emerged from these discussions with the feeling of duty accomplished. This is what he calls the soft power canadian. The fact that he represents the Canadian population when he addresses companies to ask them to improve their practices is enough, according to him, to have a lasting influence on their behavior.
At closed microphones, François-Philippe Champagne explains that the soft power is a more common practice in the United States. President Joe Biden has used it regularly since he was elected to ask the leaders of American multinationals to correct certain bad habits.
He’s done it in AI, he’s done it in the automotive sector, he’s doing it constantly in a multitude of economic sectors.
Two years later
This “power to convince” (official translation) is a practical lever for a government that does not want or cannot legislate. The American president is paralyzed by part of his government, which systematically refuses to cooperate. He has virtually no recourse other than his power of influence.
The situation in Canada is not the same. But we feel that the Liberal government is in a hurry to act. Or at least, this is the image he wants to project: see, we are acting, even faster than the normal legislative process.
Its voluntary Code on AI is a temporary measure, pending the adoption of the Law on Artificial Intelligence and Data, which is being re-read in the Commons. Except that at the rate things are going, this law, one of the first of its kind in the world, may not be adopted for another two years. It will take more time to come into force, the time to create the machinery necessary to put it in place.
One of the central elements of the bill is a new position of AI commissioner, who will ensure that Canadian companies using AI comply with the law.
This commissioner will have a colossal task before him. The Artificial Intelligence and Data Act dictates major principles that must be respected by AI, but it will be up to the commissioner to define the concrete limits and penalties, in the form of regulations that can be quickly put in place, at case by case. A bit like the CRTC does with telecommunications.
The position comes with some potential, but also limitations. A commissioner who is too close to the industry could be very lax, while a commissioner who is too strict could scare away companies specializing in AI that do their research and development in Canada.
Alone in the world ?
In the meantime, the question remains: what is this power of influence of the Canadian government in fact? Can we imagine that the AI industry will comply with its demands?
We see it elsewhere in tech, where the influence of the Canadian government is relatively weak, if not downright negligible. If Ottawa had real power of persuasion, Google and Meta would probably not behave the same way in the face of its current and future laws. Including the Online News Act, which, for the moment, is hurting Canadian companies much more than these two American giants.
Because, from their point of view, Canada’s whims do not prevent business from continuing to move forward. Google launched smartphones last week that it presents as “AI in your hand.” Its main new features are based on its generative AI, called Bard.
However, these new products are available in 220 countries around the world, but not in Canada. Nor in the countries of the European Union which also plan to legislate on AI in the near future.
On online news, Meta doesn’t seem to want to go back. If Meta bows to Canada, it could then have to bow to California, Europe, and other governments who also want to legislate on AI and GAFAM.
And if François-Philippe Champagne can boast of representing 40 million Canadians, this influence is rather slim compared to 330 million Americans and 450 million Europeans. His power of influence could turn against him if Meta, Google and others manage to discourage other governments on the planet from imitating Canada