Canada and Europe are thinking about it. The United States did it. Quebec is studying the question: what would be a good legislative framework for artificial intelligence (AI)? She will be debated Thursday in Montreal during a public forum organized by the Quebec Innovation Council. Chief innovator Luc Sirois expects a unique event in the world.
“We don’t see a collective approach like that of Quebec elsewhere,” said Mr. Sirois from the outset in an interview with The duty. Thursday, during its Public Forum on the supervision of artificial intelligence, the Quebec Innovation Council will present the findings and recommendations that emerge from all these consultations and contributions. The main highlight is the importance of gaining public trust. “Nearly 300 experts participated. We also appealed to the public — 420 people contributed. »
“What is emerging is the importance of being at the front of the parade — of continuing to invest in research and development in AI everyone agrees. Then, you have to do it responsibly — include verification tools that ensure that the tools and data are reliable and trustworthy. »
It goes quickly
Leading the way in AI is easier said than done. Thanks mainly to the importance of university and private research in AI carried out in Montreal, Quebec has a surprisingly central role in the emergence of this technology. However, the accelerated emergence of generative AIs like those of OpenAI and Google have completely changed the portrait of the industry.
“Generative AI is quickly coming to Tabarnouche! exclaims Luc Sirois. More than with the Internet and social networks, the education, work and health sectors must evolve. This is not a new theme, but it is a virtually infeasible challenge to keep pace with AI. We are forced to invent new ways of doing things that we didn’t have before. And we must provide tools to citizens so that they know who they are dealing with, who they can trust. »
This ability to detect the presence of AI or the nature of content generated by artificial intelligence is another recurring element among the comments received on the subject by the Quebec Innovation Council, explains its main manager.
Here again, the challenge is significant. There is consensus on the desire of governments, specialized companies and researchers to force their computerized creatures to sign their works so that the source can be found. As the influential Montreal researcher Yoshua Bengio has mentioned several times in recent months, the threat lies more with AI applications which will be diverted from their initial role by people or organizations with malicious intentions.
We have already seen Facebook being manipulated by foreign groups to influence electoral results in several democracies. Every day we see the Internet network being undermined by pirates whose sole objective is to empty the bank accounts of Internet users all over the planet.
Yes, there are criminals, but that doesn’t mean we shouldn’t establish a framework
It would be illusory to believe that AI will not suffer the same fate. “Everyone shares this opinion. The threat comes from rebellious countries. What can we do ? asks Luc Sirois. I find that a bit defeatist — yes, there are criminals, but that doesn’t mean that we shouldn’t establish a framework. »
AI in all its forms
If the high level of participation in the Quebec AI supervision project is particular enough to intrigue internationally, it is because almost all Western governments currently have this file on their drawing board. “The French Minister of Digital Contacted us to better understand, Sweden, Austria, Germany are all intrigued. What we do in Quebec is quite unique. »
It is in any case very different from what the American President, Joe Biden, did last Monday, who put an end to the procrastination and suddenly imposed by decree his own legislative framework to limit possible slippage in the use of the AI on American soil.
The presidential order draws inspiration from proposals put forward elsewhere, notably in the artificial intelligence and data bill that the Canadian government hopes to adopt soon. For example, the United States will create an AI Safety and Security Council, which will assess the threat the technology poses to the country’s security and the privacy of its citizens. Canada is also considering creating a council, an organization that could roughly do for AI what the CRTC does for TV and radio.
Obviously, the United States has an influence on the technology industry that can only make Quebec dream. But the province can take inspiration from it, says Luc Sirois. “Creators of new technologies must also comply with the law, and if they do not, they will be priced out of the market. The example of drugs is a good one: hundreds of billions of dollars are invested in research and development even though it is extremely regulated — and those who cheat end up getting banned or they end up in prison. »