Mr. Speaker, I am so glad we are debating the motion today, because today is the one-year anniversary of the release of a generational, society-changing technology, and that, of course, is ChatGPT.
I think I was the first legislator in the world to stand up in a parliament and say “Hey guys, have a look at this.” Since that time, over 80% of Fortune 500 companies are now integrating ChatGPT technology. Legislators around the world are trying to deal with the vast societal implications of the release of the technology. Also, the world is trying to grapple with the fact that the technology was released into the world without any sort of comprehensive regulations around the development of large language models and the large-scale deployment of this type of technology, and then without a lot of thought to use.
The other thing is that, in the last year, we have had to build, internationally, parliamentary capacity for legislators both to understand the technical aspects of how artificial intelligence has the capacity to impact our society and to try to look at how our regulatory systems can meet the challenge. Our systems are notoriously non-nimble and slow, and this is why there should be no partisan divide on the fact that the motion should pass. The Artificial Intelligence and Data Act should be hived off. Without offence to the government, the Artificial Intelligence and Data Act was developed about a year or a year and a half prior to the deployment of ChatGPT. It is like developing regulations for scribes, putting them forward and debating regulations for scribes, after the printing press was deployed around the world. That is really the transformational change we are dealing with.
Taking a partisan hat off, civil society, academia and industry all need to be consulted, and we need to go back to the drawing board on a lot of places in the bill. As my colleague for Bay of Quinte talked about, this is something that the Standing Committee on Industry has heard over and over again from every witness it has had. However, it is also important to split the bill to give the government an opportunity to better coordinate with other jurisdictions around the world that are trading partners with Canada, are already well ahead of us and are close to passing their own artificial intelligence regulations. For example, the European Union is already well down the path, as is the United States. Frankly, we also need to include the global south in the conversation. We need to be working with the global south, as it impacts the global south.
I am pleased to announce that the Canadian Group of the Inter-Parliamentary Union successfully passed a motion at the IPU's last assembly in Rwanda a few week ago, and that Canada will be a co-rapporteur on the Standing Committee on Human Rights to develop a motion specifically to do this, to build capacity for parliamentarians around the world to learn about the impacts of artificial intelligence, to take that back to their respective political parties and legislatures and to ensure that all voices are being heard. This is one of the reasons we have to pass the motion before us today.
I want to echo what my colleague for Bay of Quinte said. I think there is one testimony that sums up why AIDA has to be hived off so the privacy components of Bill C-27 can proceed. I will read from Barry Sookman, senior counsel at McCarthy Tétrault, who argued that AIDA fails to adequately shield the public from potential risks associated with high-impact AI systems. He also said that the centralized power that was envisioned in this pre-ChatGPT legislation undermines the structure of parliamentary sovereignty, adding that “AIDA sets a dangerous precedent”.
These are Canadian legal experts who have made the argument that the implications of implementing under-considered legislation on an issue as impactful as artificial intelligence is extremely risky. Any flaw in AI regulation could affect millions of people, exposing companies to class action lawsuits of historic proportions. It could also expose the public, our constituents, to risks, because we have not thought this through. This issue is so huge. It has the capacity for so much societal transformation that the bill must be hived off. Members from all political parties need to be engaged in robust, fulsome debate with all aspects of Canadian society. They need to think about this in three silos.
The first is the way that artificial intelligence is developed. It has come to light over the last year that ChatGPT was developed using extremely low-paid labour in the global south. These low-paid labourers were exposed to violent child pornography imagery to help train the large language model. There are no global regulations or standards around this. That needs to change.
The other thing there is really no global standards for, certainly not in Canada, is the protection of intellectual property when it comes to training large language model systems. This is highly problematic. We are already seeing precedent-setting legal cases coming forward in other jurisdictions, which could have extreme impacts on Canadian businesses, the ownership of IP and also how we promulgate and respect our trade agreements with other partners. A lot of our trade agreements did not consider artificial intelligence.
The second is the development of artificial intelligence. The fact that ChatGPT was released on the public, where a hundred million users are using this on such a regular basis without thought to what that means, is like releasing a pharmaceutical onto the public with no clinical trials, with no data. As a country, we need to think about how we research these products, how we allow research and innovation, but also we need to ensure that the societal impacts are thought about in an ethical framework prior to deployment.
The last thing is that I want to encourage colleagues to join the parliamentary caucus on emerging technology. I have colleagues who are sitting here today who I know have such a heart for these issues. This debate has been in a non-partisan manner. It has been collaborative and it has been great.
Just briefly, there is the impact of AI on democracy with deepfakes, with the spread of information, and on labour and the disruption of labour. Will we see AI replacement workers? Is that being considered in any legislation? These things need to be considered in an AI regulatory framework: cross-jurisdictional issues, the issues of human rights, the issues of autonomous weapons.
I could go on and on, because the bill needs a separate vote. Probably, the government needs an opportunity to go back to the drawing board to internalize the situation. This needs to happen now, though. I think that this is a no-brainer. I think there is a lot of consensus in the House of Commons for that. I would like to see agreement in the House on this matter and it be put to a vote.
With that, I move:
That the debate be now adjourned.