INDU Committee Meeting

    Good morning, everyone. I hope you had a good weekend.

     I hope you were able to get home for a little bit and spend some time with family and constituents.

    We are joined today by Minister Solomon, who is here to talk to us about artificial intelligence here in Canada and a variety of things in relation to that file.

    Minister, as you know, you'll be afforded the opportunity to make some introductory remarks, followed by lines of questioning from our colleagues around the table.

    Accompanying the minister is associate deputy minister Mark Schaan, who's been here before in this capacity with Minister Joly.

    Welcome back, sir.

    With that, Minister Solomon, the floor is yours for some introductory remarks, sir.

     Thank you, honourable Chair.

    It's great to see you, members of the committee. Thank you for having me.

    Hello.

    As you know, I'm joined by officials from my department, including Mark Schaan, my deputy minister, and others. I appreciate the opportunity to be here today.

    On June 4, the Prime Minister and I launched the AI for all strategy.

    It's Canada's national AI strategy. It's based on three principles: trust, opportunity and sovereign control.

    AI for all means no Canadian—wherever they live, whatever their economic position—should be left behind by this technology. It means AI must serve Canadians, not the other way around. It's AI that helps make health care better and AI that makes public service faster and more effective.

    In Halifax, Dr. Robert Chen, who is a Canadian doctor, is helping to build an AI-powered stethoscope to screen heart murmurs in babies and children. That means faster answers for anxious parents, fewer unnecessary specialist referrals and shorter wait times for children who need care the most. It's a Canadian doctor and Canadian innovation.

    In Saskatchewan, Jonathan Adams farms with his parents and his brother. With Croptimistic's AI soil mapping, they treat each part of their field differently, using fertilizer where it helps and where it needs more, and using less where they don't need it, to improve yields with less waste. It's a Canadian innovation and Canadian workers.

    AI is helping workers, farmers, entrepreneurs, researchers, students and small businesses improve and strengthen their lives. Right now, they're doing it with technology that basically didn't exist two, three, four or five years ago.

    We know that while some Canadians are optimistic about AI and what it can do, others have real concerns. There are real concerns about job loss. There are real concerns about privacy, personal information, children's information, our democracy and our sovereignty. We take those concerns very seriously.

    We should be open to the massive upside of opportunities for AI, but very candid about the real concerns. In a sense, Chair, there are two teams here. There is team pompom, which is the AI cheerleaders who think any regulation might be bad for the economy. Then, there's team pitchfork, which believes we should just stop it altogether—stop building any of the infrastructure and stop investing in Canadian innovation.

    Our government is not team pompom or team pitchfork. We are team pragmatic. This is a pragmatic, pro-worker, pro-innovation, pro-safety and pro-trust strategy. We're building reliable, safe, human-centred Canadian AI that serves Canadians.

    We have a choice here. We can let others build the AI infrastructure and then rent it from them. We can let someone else create the jobs, control the innovation and set the rules, and we'll have to follow them, or we can build here. We can set our own rules here. We can create the jobs here. We can keep the innovation here.

    That's the choice. We've made ours. That is our plan: AI for all.

    It begins with the first pillar, which is trust. We made a promise to Canadians that we will protect your children, we will protect your privacy and we will protect your personal information.

     Within a week of launching our AI strategy, we acted on that promise. Last week, the government introduced the safe social media act to make digital services safer for children and hold platforms accountable for the risks their services create.

    Today, we tabled the protecting privacy and consumer data act. This bill gives Canadians more control over their personal information. It protects children's data by treating it sensitively. It gives Canadians a right to deletion, so they can take down deepfakes. It aims to combat surveillance pricing to avoid price gouging. It brings more transparency to those automated decisions. When an algorithm or a computer makes a decision about your mortgage or your job, there should be transparency. There will be with this law. It gives a new regulator real power to hold bad actors accountable.

    Together, these bills do what we said we would do. They protect Canadians. They build trust. They create the conditions for responsible adoption.

    It's a simple approach: protect Canadians and build confidence for the future.

     Trust and protection are not separate from AI adoption. Trust is what will make adoption possible, but AI for all is not just about trust. It's about opportunity. It's about sovereign control. Opportunity means helping Canadian workers gain the skills to use AI confidently.

    Opportunity means that Canadian businesses are helped to adopt AI so they can become more competitive.

     AI for all will create and support up to 90,000 AI opportunities for young Canadians. We will create those and support those. It includes a national AI literacy push for free, so all Canadians have access to AI literacy and are not left behind by this shift.

    We want to increase the adoption of AI by Canadian businesses from about 12% today to 60% by 2034.

     Our strategy includes AI missions targeting health especially, so that Canadians can see how AI can reduce paperwork for doctors, improve patient flow in hospitals and emergency rooms, support clinicians, support drug discovery and deliver better care.

     It will expand our safety capabilities around AI. Because innovation has to be matched responsibly, we're doubling the budget for the Canadian AI safety institute, and it strengthens Canada's sovereign infrastructure so that systems that power our economy are not entirely dependent on decisions made elsewhere but here.

    AI for all is artificial intelligence that Canadians can trust, can use and can benefit from.

    That's why leaders across the AI ecosystem have welcomed the strategy. As Yoshua Bengio, the Turing prize winner and maybe the most famous scientist in the world, who is here in Canada, said when we launched, “The Canadian AI strategy...advocates for the development of technology that is safe, ethical, trustworthy, and that benefits society as a whole” and “This is essential to preserve Canada's technological sovereignty”.

     This is how Canada will lead: by protecting Canadians, creating opportunities and building the technology of the future under Canadian sovereign control.

    Chair, thank you so much. I look forward to the committee's questions.

     Thank you very much, Minister. We'll go right into that.

     Ms. Dancho, the floor is yours for six minutes.

     Thank you, Mr. Chair.

     Thank you, Minister, for being with us today.

    As I'm sure you may be familiar with, Canada's Communications Security Establishment has warned that AI is, of course, transforming the cybersecurity threat faced by Canada, and leading cybersecurity experts have argued that governments need to invest specifically in cybersecurity talent that is really capable of defending against AI-enabled attacks on and threats to Canadian critical infrastructure.

     Certainly, what we've heard is that AI supercharges cybersecurity attacks and cyberwarfare, as I'm sure you're familiar with. Really, it's representing a very grave threat to the individual, to business and, again, to our critical infrastructure, which, as I have heard over the years from cybersecurity experts, both internally and externally, is in fact the most likely and immediate threat that Canada faces—a cyber-attack on our critical infrastructure such as, for example, the Northeast Power grid in the middle of winter.

     I was surprised to see that your AI strategy did not focus on cybersecurity. It was not a dedicated pillar. It wasn't emphasized. Can you explain why that is?

     Thank you for the question.

     We take cybersecurity risks absolutely seriously. I want to thank the member for raising an absolutely critical question. As the member knows, in our strategy, first and foremost we doubled the budget of the Canadian AI safety institute. We work closely with Canada's cyber-safety institute on threats that are on the horizon: threats from AI models and cybersecurity threats.

     Shared Services Canada, for our critical infrastructure, has multiple layers of cybersecurity defences—cyber-walls, network defences, anti-denial of service measures, antivirus measures and virtual private networking—but one of the key protections is investing in Canadian sovereign AI and Canadian infrastructure.

    If I may interject, you did raise the AI safety institute. Again, I do appreciate that the strategy does have an AI safety focus, but of course that's not the same thing as cyber-defence. AI safety is, of course, focused on things like deepfakes, and I know that the AI safety institute focuses on model evaluation, governance, transparency and responsible deployment—all very important—but that's not the cybersecurity defence emphasis that I'm asking about.

     Specifically, even if it did, out of the $2.3 billion you committed, only $50 million is going towards the safety institute, which is about 2%. Again, that's not solely dedicated to cyber-defence. In fact, that's not the focus of the safety institute. Can you identify a specific funding stream in your AI strategy announced the other day that is specific to cyber-defence?

     Again, thanks for the question. It's a really important question. Obviously, we take the safety of Canadians, their privacy and cybersecurity threats really seriously.

    The defence industrial strategy, which we have also made public, makes major investments in high-security organizations and in our most important cybersecurity organizations, as you know, through CSE and other organizations. Our defence industrial strategy has made significant investments in that.

    I want Canadians to know, Chair, that these strategies are working side by side. On the defence industrial strategy, we are working very closely with the AI strategy to make sure that all elements of defence and security are taken care of.

     Thank you very much.

    Again, you have a lot of emphasis in this strategy on the opportunity, which I think we've all heard. There's great opportunity with AI, but again, we have heard at this committee and others that the cyber-threat is very real, very rapidly evolving and growing. Again, these cyber-attacks can now act and evolve on their own, which is a great difference from hackers in their basements or bad government actors, for example.

    Unless you can correct me, there isn't a dedicated talent stream for cyber-defence, but there are for the positive aspects of this, which I find concerning. Is there a talent stream for cyber-defence in your strategy? I couldn't find one, but maybe I'm wrong.

    Chair, the member will find, in the defence industrial strategy, a major focus on security and cybersecurity. I would also just say a couple of things.

    We are working very closely with Canada's cybersecurity centre on some of the models that the member is right to bring up, for example, Anthropic's Mythos model. This is a very powerful large language model that Anthropic created. There's concern that it could have cybersecurity implications.

    They created something called project glasswing to allow certain governments and organizations within those governments to get access to Mythos. For Canada, I am happy to say we've worked closely. We have access, and we are part of project glasswing to make sure that our critical infrastructures.... I will assure you that I'm in close touch with my U.S. counterparts on that. I've had meetings with our national cybersecurity and our top security officials with the U.S. to make sure that we are co-operating on fortress North America.

     I appreciate that you are taking meetings on this. It sounds like you're having some discussions, but again, you're the first AI minister this country has ever had. I do feel that you have a responsibility to make this a primary pillar of your AI strategy, given the gravity of the situation regarding cyber-attacks and their evolution by AI-enabled infrastructure.

    I think we're out of time here, but I would just urge you to reconsider your strategy and ensure that if you evolve it or bring forward a 2.0, cybersecurity is a main pillar. It's a grave enough threat and a serious enough matter that the first AI minister in this country should have it as a primary pillar. It is concerning that it was not a primary focus in your strategy.

     Cybersecurity remains absolutely vital, and security is vital. One of the reasons we are investing so much in Canadian sovereign AI is to make sure that we have control over our systems, that Canada's personal information and our sensitive workloads are held by Canadians under Canadian law. This is vital and central to our strategy to build sovereign AI for the very reason that the member rightly brings up: to make sure that we are safe and that we have control.

     Okay.

    Thank you, Ms. Dancho.

    Mr. Ntumba, I understand you're going to share your time with Ms. O'Rourke. I'll let you decide how you want to split it up.

    You have the floor for six minutes.

    Thank you, Mr. Chair.

    Good afternoon, minister.

    Recently, the government introduced a bill on youth and social media. Next, there was the digital strategy, and today, you introduced another bill.

    For my part, I would like to take a much more youth-focused approach. Deepfakes are becoming increasingly realistic and accessible, particularly on platforms that young people use, posing significant risks of disinformation, cyberbullying and reputational harm.

    What concrete steps are you taking to protect young people from the creation and dissemination of deepfakes, and how do you plan to hold the platforms in question accountable?

    Thank you for that very important question.

    I'm going to have to switch to English. I used slightly simpler sentences in French, but trust is very important.

     Safety is also important. Our strategy focuses on trust, opportunity and sovereignty.

    It makes sure, first and foremost, that we are investing in young people. As I said, we are creating a national AI literacy program. Every single Canadian will have access to literacy. We will be providing every young person and post-secondary student with a reliable, safe AI tool they can use, so they have access to these tools. That's very important.

    Third, we're going to be creating jobs. Working with Canada summer jobs and Mitacs, we will be supporting 90,000 job placements so young people have paid job experiences in the workforce and skills for the economy of the future. In the same way—per the spring economic update—we are training 100,000 skilled workers in Red Seal trades. This way, in the technological and the digital economy, we are training 90,000.

    You asked specifically about deepfakes and trust. I am very proud that we just tabled legislation today on protecting privacy and consumer data. One of the elements of that would give citizens the right to combat deepfakes. It's about having a right to deletion. If there is something posted that they don't want up, they can demand that it be deleted or taken down. This is a very important step that works hand in hand with the legislation we tabled last week on safe social media. That's absolutely important.

    Again, there are other ways we are investing in young people, like CanCode for young people. I know the honourable member has some young kids. I have kids. It's about skills training from K to 12. Mine are a little older, but not yours. They will benefit from that CanCode training. We are investing in skills, in literacy and in giving them power with a right to deletion, and we're protecting them from the harms of social media with our safe social media act, restricting access to kids under 16.

    Thank you very much.

     Thank you, Minister, and welcome.

    I have a couple of questions. I attended a session with ParlAmericas, where people from all over the world were talking about data being the new oil. People are concerned about the sovereignty of their own personal data and their ability to control it. I'm wondering how this new piece of legislation allows Canadians to feel confident that their own data is secure and that they have control over it.

    Then, in a broader frame, the Munk school talked about different levels of risk and how we need to protect different types of data differently. I'm wondering if you could speak to that.

     Thank you. It's an excellent question, and I appreciate it.

    Data sovereignty is core to this strategy. We want to make sure Canadians' most sensitive data are protected here in Canada under Canadian law, and that it's also not subject to the coercion of other laws. We need the infrastructure to build that. We need to build the infrastructure to make sure we have all aspects of what they call the “ecosystem”, whether it's about building data centres or building them the right way. There will be questions about sustainability—water use like closed-loop water or making sure we're reusing the heat. We're being very transparent about it and building with our values. All the way up, we need Canadian compute. Computers are the fuel of the innovation that companies need so they can buy Canadian. That's very important.

    I often say that sovereignty is not solitude. Canadians have access to all the tools of innovation from countries around the world. Many of those countries are here. They employ Canadians. They employ engineers. They've been here in communities for a long time. We welcome that, but Canadians also want a sovereign option. That's why we're investing.

    Let me read how we're building that. It's not only by investing in SMEs. This is one of the largest investments in technology innovation for small and medium-sized businesses and Canadian innovators in Canadian history. We have $1.5 billion through the BDC, the Business Development Bank of Canada, and their LIFT program. We have a $500-million tech growth fund to invest in innovators and reward people starting businesses here. We'll encourage them to do that. We have another $500 million that will be invested through regional development agencies because there's great innovation on the ground. Then we have about $700 million in what's called the compute access fund, so small companies are incentivized and subsidized to get compute and to buy compute from Canadians. That's really important. We're investing in our researchers as well.

    I'll read you a quote from the CEO of eStruxture, Todd Coleman. He said, “The federal government's AI strategy is a positive step that recognizes Canada's AI ambitions and the data economy depends on secure, resilient, Canadian-based digital infrastructure that gives businesses, governments and communities confidence in where data lives and how it gets used”. He also said, “Building value for Canada needs to be done with Canadian values”.

    We're really happy to see that kind of support.

     Thank you.

     Thank you, Ms. O'Rourke.

    Mr. Ste‑Marie, you have the floor for six minutes.

    Thank you, Mr. Chair.

    Good afternoon, minister and deputy minister. Thank you for being here.

    I find your policy very interesting when it comes to promoting the development of artificial intelligence and its adoption by businesses, accelerating its use by government and fostering productivity. As far as oversight is concerned though, there is nothing.

    Do you intend to introduce a bill on artificial intelligence that will accompany your strategy, similar to what was in the part 3 of Bill C‑27?

     I'm just trying to make sure that I'm answering the right question here.

    I want to inform the honourable member that the AI for all strategy is structured around trust, opportunity and sovereign control. I want Canadians, as well as the honourable member, to know that there's a page of clear results of what we expect if we follow the investments in this strategy, such as creating, in five years, 250,000 new jobs; driving adoption rates from 12% to 50-plus per cent; creating 90,000 jobs for kids; an AI literacy strategy; and an AI tool. There's a lot of framework there, and there are almost 50 action items. I'm just trying to be clear.

    Thank you.

    There's an enormous amount of detail.

    Minister, with all due respect, there are a lot of things to develop the industry, but when it comes to oversight, for example, how do you intend to protect creators whose works can be used by artificial intelligence models when your strategy doesn't address the issue at all?

     In fact, just on the creatives, there is.... Again, I appreciate that for the member, it's a long strategy. We have $50 million dedicated immediately to the creative technology program to support Canadian creators using AI on their own terms, which is very important. I've also had a two-day summit with culture leaders and stakeholders, and we are setting up an AI culture advisory group—both myself and the minister of heritage—which is absolutely critical.

    We are investing in, obviously, our culture makers in a very profound way, and we're pretty clear about that in the strategy.

    In terms of protections—as the honourable member asked about—again, within 10 days of the launch of our strategy, we have launched comprehensive safe social media legislation that restricts access to social media for kids 16 and under, among many things—

    Thank you.

    —including capturing AI chatbots, and today I tabled the protecting privacy and consumer data act.

    I understand. You said that in your previous answers. Thank you. In any case, we would like an answer regarding the protection of works.

    The same goes for citizens' data. When the government uses artificial intelligence software to process people's data, what steps will be taken to ensure that the data isn't stolen by artificial intelligence software, that it doesn't end up elsewhere and that it is properly protected?

    Thank you for the question.

    It's a really important question about protecting Canadians' data. I do appreciate that it might have been.... Literally just before I got here, we tabled the protecting privacy and consumer data act that gives extensive detail about the obligations that all companies will have with regard to protecting children's information and protecting consumers' information: how they have to deal with a chain of custody, what responsibilities they have when they're collecting it, and whether they're collecting it for beneficial and reasonable purposes.

     It's a super important question, and I'm glad you've asked it. That is exactly why right after the national AI strategy, AI for all, was tabled, we tabled this new legislation to protect and to make sure that Canadians' data is protected.

    Your other question is about, maybe, training and data mining, what's called TDM. There are real concerns about copyright. My colleague Marc Miller is seized with that. There are obviously some ongoing court cases on that. However, the principle of making sure that Canadians are properly compensated for their work remains in place.

    Obviously, we have a lot of concerns about protecting the information of citizens and individuals when artificial intelligence software is used by government. If the government uses that kind of software, how can we ensure that such data is protected?

    Today, the Quebec minister of labour tabled his guide on the use of artificial intelligence in business. It outlines for example what has to be done to prevent workers from being spied on by artificial intelligence software and to ensure that it helps increase productivity and not replace workers.

    Does your strategy include provisions to protect workers who use artificial intelligence software?

     First of all, jobs are a critical part of the strategy on the training and the education.

    The member asked about governments. The proper regulations around how AI is used by government, the Treasury Board regulations, are absolutely in force. There's an AI registry that's fully transparent. There are about 413 uses of AI within the government, and there is an AI registry that's very transparent. You can look at that. There are regulations that are very thorough in terms of how AI is used, what humans are in the loop, what kind of transparency and what kinds of disclosures there are any time the government is using AI. That's quite rigorous.

    In our strategy, we are making sure that workers are a part of it. I just want to say I meet regularly with labour leaders. Also, we have struck an AI labour advisory council. It will be meeting with me at a minimum of twice a year. We've already met with labour leaders about it to make sure that policies around the development of AI through this strategy and these action items are pro-worker and to make sure that we have workers in mind.

    Thank you very much.

    Thank you, Mr. Ste‑Marie.

    Mr. Guglielmin, the floor is yours for five minutes, sir.

     Thank you, Chair.

    Thank you, Minister, for being here today.

     Minister, your AI document commits to ensuring that every post-secondary student has a trusted AI agent. Who decides what a trusted AI agent is and what body gets to make that determination?

     This is a really important aspect of the strategy. We believe that AI for all means that everybody should have access to an AI tool, and that tool should be reliable and trusted. As you know, with the safe social media act and the protecting privacy and consumer data act, we'll be setting up a digital safety commission. It's absolutely crucial that we finally have a commissioner who will oversee the responsibility of how information is treated and what is reliable and what is safe. I will just tell you that this particular item is in development. We are working closely with our national AI institute—

     Minister, I'm going to interrupt here. Is the government going to have editorial or content authority over these agents through a commissioner who's being set up?

    There are two parts to your question and it's a good one. The first part is, what tool will we use? We'll be consulting with our national AI institute, the Canadian AI safety institute. We will absolutely make sure that any tool we use complies not only with the safe social media act but with making sure that kids' information—in this case they will be post-secondary people—is treated and protected according to the protecting privacy and consumer data act. Absolutely, that's where we're at. We're finally going to have a regulator. We're going to have enforcement, and we're going to have clear lines for this to make sure that whoever provides it will have to abide by those.

    You can understand why there would be some concerns. We saw the government use Bill C-11 to give the federal regulator authority over what Canadians see online. Bill C-18 reduced the news availability despite its promise to expand it. How is this government-administered AI agent scheme different in principle from these other interventions?

    I think the first principle here is equal access to the tools. You and I both know that.... When I raised my kids there were some kids who had more access to tools and others didn't. We had a digital divide. We had a digital divide in terms of remote communities, like rural communities. First and foremost, this is about making sure everyone has an opportunity—

     Minister, respectfully—

    —to have a tool to use if they're at a CEGEP or a university or a college.

     Respectfully, that's not the question. A lot of these AI agent chatbots already exist. They're free and they're open source. How does a government-directed one that's trusted...? How is that something that is enhancing access when it's already accessible?

    First of all, we welcome that there's lots of accessibility. We want to make sure that all young people have access to something that is safe and reliable.

    By the way, some of those things can be expensive. There's a lot of money on compute, and eventually there are charges. There are free models, and then there are not. Some of those models may not be safe for data, so we have to make sure they're safe and compliant.

    You mentioned open source. We have to make sure of what's in an open source as well, which is an alternative. By the way, we're interested in open source as well, to make sure that it's reliable and safe. If post-secondary students are using something, I think it behooves us all to make sure that it's reliable, secure and safe.

    Essentially, you want to determine what the AI agents are spitting back to the post-secondary students who use them.

    I want to get another quick question in here.

    A new Angus Reid poll shows that 68% of Canadians would oppose large data centres near their homes, and we know that having these data centres is important for the development of AI and AI sovereignty. Your document acknowledges energy and water concerns, but it lacks enforcement mechanisms. Here at this committee, we heard testimony that Canadian data centres operate with no mandatory disclosure regime for energy and water use. Is this something you're willing to consider implementing and putting a lot of onus on?

    I ask because companies make commitments that they're not going to use an excessive amount of energy or water and increase the costs to local communities, but there's no enforcement mechanism and no reporting required. Is this something that you guys would be looking to implement?

    First of all, you're asking exactly the right question about transparency around data centres. We need to build Canadian infrastructure. Either we build it or we rent it back from someone else who is going to build it, and Canadians will be charged for it. However, we have to build it right.

    Quickly, there are three things.

    One is that power allocation is the purview of all the provinces. The provincial regulators allocate it. It's not the federal government that allocates it. I have talked to all of my counterparts to make sure that when they're allocating power—and this is provincial—they should make sure that ratepayers are not paying more. That's a provincial jurisdiction.

    When we enter, we have an enabling large sovereign infrastructure fund to use when we are looking at data centres. We have criteria on what we will partner in. You might notice that we've begun a partnership with Telus in British Columbia for three data centres. They're hydroelectric, with a closed-loop water system, which means they reuse the water. The heat is going to be used to heat 150,000 homes. As they say, they want to use every molecule twice. It's transparent, so you're right about all of that, and it's not putting pressure on ratepayers so ratepayers are not paying more.

    You're asking the exactly the right questions. We are making sure, from our end, that if the Canadian government is supporting those data centres, they're done right.

     Thank you.

     Thank you, Mr. Guglielmin.

     Mr. Bains, I understand you might be splitting time with Mr. Ma. As usual, I'll leave it for you two gentlemen to figure that out. You'll have five minutes.

     Thank you, Mr. Chair.

    Thank you, Minister, for your work on this strategy and for working closely with British Columbians and leaders out there on the AI strategy.

    I'd like you to expand a bit on public safety and the trust pillars you talked about. Among British Columbians and all Canadians, there are skeptics out there. Specifically, we witnessed a tragic incident in Tumbler Ridge, British Columbia, that shook a community and our nation. Numerous court cases against AI providers allege that their products have counselled youth to commit harm to themselves or others, so can you talk a bit about the accountability measures that are in place?

    In this instance, it was ChatGPT that the youth confided in. She was planning to harm others, but the owners, ChatGPT, didn't act in advance. Could you please expand a bit on that?

     I appreciate the question.

    Let's just say there is nobody in this room whose heart doesn't go out to the families in Tumbler Ridge. I know this is not a partisan issue. Every one of us here recognizes that we have to protect Canadians and protect kids. Tumbler Ridge is a horrific tragedy that is ongoing for the families and the community, and we took action.

     Immediately after we found out the shooter's behaviour had been flagged eight months before by the technical side of OpenAI—in other words, the AI itself brought it up—and it was not escalated, I not only summoned the entire senior team to Ottawa but I spoke to the CEO, Sam Altman. I demanded they work with the Canadian AI safety institute to review all other past incidents to make sure they had escalated them with police and that they have a presence here in Canada with Canadian local law enforcement as opposed to just talking to U.S. enforcement, and that they have mental health experts. All options were on the table.

     The Canadian AI safety institute is still working very closely on assessing. In the meantime, we have taken action. We have tabled legislation called the safe social media act. Not only will it restrict access for young people under 16 to social media, but it will include AI chatbots that have a duty to keep children safe, a duty to be responsible and a duty by law to be transparent about that. We are one of the first countries in the world to have any regulation at all on these new technologies called chatbots. We have taken action immediately on that.

     That dovetails with what we've just tabled, which is the protecting privacy and consumer data act. That gives all companies, including AI companies, an obligation by law to protect children's information, and it gives citizens the right to see where abuses of information might take place. We have taken protective steps, but we've also doubled the budget of the Canadian AI safety institute. We know that it has to be more robust to make these kinds of assessments.

    Thanks.

     I want to cede the rest of my time to Mr. Ma.

    Mr. Ma, you will have about 75 seconds.

    Thank you.

     Minister, it's claimed that the AI for all strategy is light on details when it comes to jobs, but the strategy sets out concrete targets of up to 90,000 AI-related job and placement opportunities for young Canadians, up to 250,000 new AI-relevant jobs by 2031 and a major increase in business AI adoption.

    Can you walk the committee through how the strategy turns AI from a source of anxiety into a source of opportunity for Canadian workers, students, SMEs and young people entering the workforce?

     I know I can come back to it.

     Just to remind people, there are 150,000 people currently working in the AI industry and 800,000 people in the digital innovation space. It's the fastest-growing sector of the economy. We made sure that we're creating jobs. We will actively be supporting those job placements for 90,000 young people. We believe the OECD data shows that if we invest in SMEs as we are, we'll create 250,000 new jobs in the next five years. The investment in SMEs, in innovation, in jobs and in skills training is an essential pillar to make sure of, and the details are very clear in the plan.

    Thank you.

    Thank you, Mr. Ma.

    Mr. Ste‑Marie, you have the floor for two and a half minutes.

    Thank you, Mr. Chair.

    Minister, I'm going to raise three points right now about your strategy, and then you can respond.

    First, the strategy makes no reference to a possible international treaty, even though Canada sponsored a draft resolution to that effect at the United Nations in 2024.

    Second, in terms of partnerships with Europe, the policy refers to market development and supplier diversification, but not regulatory harmonization. Why not?

    Finally, your strategy does not mention ethical considerations in the development of generative artificial intelligence. Why not?

    Thank you for those good questions.

    There's a lot of information in the strategy about our alliance with other countries.

     We refer very strongly, for example, to the sovereign technology alliance that we've built with Germany. This is a critical alliance that has already led to one of the most important Canadian companies, Cohere, a large language model, partnering and merging with a German company called Aleph Alpha, giving Cohere sovereign presence in Europe and making it a legitimate competitor to any other large language model.

    That sovereign alliance is really important, by the way. There are only four countries in the world that have large language models: the U.S., China, France and, now, Canada. That alliance has really helped. We have digital alliances with the EU. We have a technology alliance with India and Australia, as well as with the U.K. I will observe today, Chair, that the U.K. has followed Canada in age-restricting social media. Today, they followed what Canada did last week. There's lots of coordination. I was just at the G7 coordinating that. In the strategy, pillar three, which is sovereign control, has a whole section on building sovereign alliances.

    I think there was another element to your question at the end.

    I'm sorry, I forgot the last part.

    I was asking why your strategy makes no reference to ethical considerations in the development of generative AI.

    I want people to understand that we talk about AI, but there are different types of AI. There's generative AI, which is the kind of AI that people most commonly use. They type into an AI agent, and it gives them an answer. They call it generative because it's generating a response. It's a predictive machine.

    The other kind of AI is called agentic, and not to get too technical, but it does tasks for you. That's the second stage. There are two kinds of AI. Both of those, agentic and generative, in the strategy, we just call them AI. They are all going to be, of course.... Any company subject to, obviously, either the privacy regulations.... If they're using private information, that's very important. We have invested heavily in safety and ethical guidelines—we already have a voluntary code—to make sure that AI is used to serve Canadians and not the other way around.

    Thank you very much.

    Thank you, Mr. Ste‑Marie.

    Ms. DeRidder, you have five minutes.

    Thank you, Chair.

    Minister, the AI document speaks to principles but not execution. There are some action items, sure, and some job promises, but there is no execution on how to get there. When will Canadians see an operational plan that outlines concrete timelines and economic impact for the investment in required resources, such as other ministries' involvement—procurement, defence and infrastructure, for example?

    Thanks for the question. I appreciate it was a long document.

    This strategy is by far the most detailed national AI strategy in the G7. We have a full page on metrics and how we're going to get there: 60% adoption by 2034 and 12% today, and 250,000 new jobs. You asked for targets.

     My question is about how you are going to get there. There's great language about what you're going to do, but my question is this: Operationally, how are you going to get there, and how are you going to govern it?

     Again, I appreciate the question.

    There are very specific investments tied to how we are going to do it. How are we going to improve, for example, the adoption rates? One way is by setting up our most successful program, called the compute access fund, to give small and medium-sized businesses—many of which, by the way, I know are in your riding—access to compute. It is really important to subsidize access to that so they have access. That's a $700-million plan, and it builds—you asked how—on the $300-million plan that we've already rolled out.

    We know how to do it. That's a very specific program that we are building on. Again, we're building a billion-dollar supercomputer for our researchers. We've done the RFP. The decision will be made in the fall. Those criteria are out. On building sovereign data centres, the RFPs are out.

    You can see we've already done the first project with Telus. Those are all very public, and the steps are very clear. We tried to give real targets and tie real investments to them so that Canadians can track them and have accountability.

     Thank you.

    I'm going to ask a question about how the document was put together.

    Protecting Canadians and safeguarding our democracy is pillar number one, and from that stems trust. Given that this document had over 11,000 contributions to support the strategy, how did you know that this was a sample representation of Canadian companies and it wasn't overwhelmed by the contributions of foreign actors who also partook in this strategy? How did you make sure that it was wholly Canadian voices and not from American and foreign firms?

     We are really proud of the consultation. We think it's important to have consultation with Canadians when you're setting up an AI strategy—there are real concerns—and part of that is being here at committee with you.

    It wasn't just the 11,000 really detailed contributions that we got from Canadians, which was great. We also had the 28-person national AI task force. I did a series of round tables for months with entrepreneurs, Black entrepreneurs and indigenous leaders, so we had widespread consultation.

    With all due respect, though, you're not really answering the question.

    How did you make sure it was wholly Canadian input? ISED shows that some of the submissions received were from huge, U.S.-based firms, such as Google, Tesla, Amazon and IBM.

    How did you ensure that it was Canadian voices that developed this AI document?

     The question about consultation is important. By the way, we do consult with businesses that conduct business here and that employ Canadians—businesses from around the world. There are companies that invest in Canada, and we're proud. In many ridings, our foreign direct investment is really important. We want companies to come here, because we have great workers to invest in Canada, and we support that. We made sure—

    Minister, I'm not being combative, but you're not answering the question.

    I have 30 seconds left, so I want to get one more question in.

     Colleagues, here's what I'm—

    I just want—

    Yes, I understand.

    Often it's the clock we're fighting, so I'll take that out of the equation. I'm going to pause it with 25 seconds left for you, Ms. DeRidder, and I'll be generous on that 25 seconds. We're going to offer the minister 20 or 30 seconds to conclude his remarks. It will not take away from your time. I'll allow you the opportunity to pose one more question and allow him a similar amount of time to answer it.

    Minister, go ahead.

     Thanks for that, Chair.

    Again, we took consultation, not dictation. It was important that we consulted with people and that we listened to people, but we then made the strategy, AI for all, for Canadians to serve Canadians. The investments speak for themselves. This is about trust, opportunity and sovereignty. We made sure that we drew up the strategy, but we widely consulted.

     Thank you.

    Thank you for making that easy, Chair.

    Go ahead, Ms. DeRidder.

    You just did a press release on the data protection act—Bill C-36—and we haven't had any debriefing on this as the opposition. It's very new as I'm hearing about it, but the first question I'd like to ask is about the fact that it looks like you might be putting the onus on industry to protect our data. If we have no sovereign cloud, how are you going to govern that if it would then fall under the U.S. CLOUD Act?

     I appreciate that the honourable member is in a difficult circumstance because we literally just tabled the bill, and I'm happy to have any ongoing conversations. Obviously, we want your feedback. I appreciate that.

    There are two things about that. Number one, the piece of legislation, the protecting privacy and consumer data act, is very clear on the regulations for how Canadians' private information is going to be treated, and it will be setting up, as many have called for, a regulator. It's using the same regulator for the safe social media act. That regulator will have enforcement powers to keep Canadians safe and to make sure that private information is safe.

    As for the second part of your question about cross-border concerns, I think you were asking about what happens if data crosses the border. The chain of custody is really important, and any company that is.... Data travels. They will have to fill out a privacy impact assessment. The onus is not on the businesses. They have clear rules—privacy impact assessments—for how they treat information, and the regulator can ask for it if there's a complaint.

    I think for everybody it's common sense that we want to know, when we're giving our personal information to a company, how they're treating it and that it's safe when it crosses borders.

     Thank you very much, Ms. DeRidder.

     I understand that Ms. Begum and Mr. Bardeesy will be splitting their time. I'll leave it for the two of you to sort that out. You'll have five minutes between you.

     Thank you very much, Chair.

    Thank you, Minister, for being here.

     I want to start by saying congratulations, and thank you for your work in launching the strategy and for taking into consideration Canadians and how we want to be ahead at this time. Another thing I want to also commend you on is not just the three principles but the pillars, which focus on safeguarding our democracy, making sure that we're empowering Canadians and, of course, focusing on sovereignty. All these pieces make this much stronger.

     One thing I want to ask about is something we just discussed. Canada's AI for all strategy rightly recognizes that we want Canadian workers, researchers and businesses to benefit from AI. We also need the underlying infrastructure to be right here at home. That means focusing on data sovereignty, which we talked about, which brings us to the concerns around data centres. In my riding of Scarborough Southwest, the proposed expansion of the Stack data centre has raised many important questions and concerns from my constituents about electricity demand, water usage, land usage and, of course, the impact on local communities and community members.

    Minister, could you talk about how our government is ensuring that the new AI investments—and you've just announced some as well—are planned in a way that strengthens AI sovereignty and innovation while also being transparent, sustainable and beneficial to the communities that host them?

     Thanks for the question. I appreciate your work.

    Those are real questions that are coming from communities about this. This is new. Things are moving fast. That's why AI for all starts with trust. Technology moves at the speed of innovation, but citizens move at the speed of trust. That's why the first things we've done are the safe social media act and the protecting privacy and consumer data act.

     First, Canadians must know, as the Prime Minister said, that we're going to protect your children, we're going to protect your privacy and we're going to protect your personal information. That's key.

    We have to protect our democracy as well. That's why we have Bill C-25, Strong and Free Elections Act, which the transport minister Steve McKinnon tabled. It is really important to keep our elections free from interference from AI.

    That's why the Minister of Justice has a piece of legislation to criminalize the non-consensual use of sexualized deepfake images of people and to make that a criminal act. That's why we have safe social media act, and that's why now we have protecting privacy and consumer data act.

    There's a suite we're going to protect, but what about protecting our water? What about protecting our electricity rates? First and foremost, on energy use, building the economy of the future is going to require a lot of energy. We have to make sure that ratepayers do not pay for that. This is why we have a national electricity strategy to double our electricity grid by 2050. That was launched before the AI strategy for a reason. We have to invest in our energy systems to have the infrastructure to support the economy of the future.

    Then we have to make sure that when we're building, we're doing it transparently, sustainably and in a way that will create jobs. That's why, just to give you an example, when we enter into an agreement through our enabling sovereign data centres and Canadian data centres, we want to build these with Canadian companies so the Canadian jobs are here.

    The first one—the one we have done as a government—entered into an agreement in B.C. It's hydroelectric energy, which is clean, and it's closed-loop water. That's a phrase that sounds complicated, but people think that the data centres are just going to keep using water, but “closed loop” means that the water is looped around and reused all the time. There's even a great, big Canadian technology company in Montreal called Hypertec that does immersion cooling. They literally immerse the chips and the stacks in cooling. You don't need reusable water.

     Then finally, there's heat reuse. They generate a lot of heat. That heat in the north can be used for greenhouses, as it has been. Hypertec is doing that with a data centre. It can be used, as in the case in B.C., to heat 150,000 homes. Those are really important questions.

    Again, the province is the regulator. Provinces allocate power. There are many data centres, and we have to make sure that provinces also understand that ratepayers shouldn't pay for the innovation of others. They want the innovation to benefit them but not on the backs of higher power rates, and that's real.

     Thank you so much, Minister.

     Thank you.

    Thank you, Ms. Begum.

     Unfortunately, that is the time we have. I'll let the minister's colleagues give him grief for the depth of his answers, which took away from Mr. Bardeesy's ability to ask questions. That's tongue-in-cheek; it's all in jest.

    I will note, Ms. DeRidder, that you would certainly be welcome to come back—as would others—when we take a closer look at the piece of legislation that the minister tabled today. This will be the appropriate committee for the oversight of that.

     Colleagues, I appreciate, as always, the collaborative nature of this committee. There were good questions, which allowed us to elicit some more information about the work that's being done.

    Minister, thank you for taking the time to be here.

    With that, colleagues, I'll see everybody tomorrow, as we discussed earlier.

    The meeting is adjourned.