Canadian Heritage Committee on Oct. 8th, 2025

[Marc-Olivier Ducharme]

There are certain models, such as data trusts, that allow creators to deposit data into a trust that can be used according to the standards they set. People get together to create this trust. After that, the data can be used by industry or services, provided that standards are met and, often, money is paid. In Quebec, Culturepédia is one of those trusts.

For our part, we encourage the production and creation of specific data trusts, in other words, data trusts that belong to communities, whether they be aboriginal, Acadian or Franco-Manitoban. So communities can choose how their own cultural data that has been valued, either the cultural data that has existed for hundreds of years in our archives or the data from the Internet, are both preserved and used in the cultural industry. So creators have control over their own production. This is basic. They produced content, they deposited it and they were robbed. We now have a strategy to protect them: data trusts.

Data trusts are an example of shared governance and choice. People can decide how they're going to share their content. We're in a sharing economy; we share a lot of things. The Internet allows for such sharing. In fact, one of the Internet's special features is that it encourages us to share. We want to share, but we want to be compensated for that sharing. I think that's important. It's important to protect the value that Canadian artists produce.

[Bienvenu-Olivier Ntumba, Liberal, Mont-Saint-Bruno—L’Acadie, QC]

Thank you very much, Mr. Ducharme.

Mr. Rogers, along the same lines as the question put to Mr. Ducharme, who is asking for regulations in this industry, how do you see regulations being implemented?

[Patrick Rogers]

I would say two things. The first is that the most important part is to enforce the Copyright Act, as is. The second part, in terms of looking for something new to do, is to enforce transparency laws. In order to understand what has been ingested, in order to enforce that Copyright Act, we will need increased transparency.

As I said in my opening statement, the nerds freak out when we talk about this and say that the data is too big, too expansive, and that our poor human minds could never understand it, but again, AI should be able to build a list of what it's ingesting.

[Bienvenu-Olivier Ntumba, Liberal, Mont-Saint-Bruno—L’Acadie, QC]

Ms. McGuffin, as parliamentarians, our expertise in the musical field is quite limited. Can you explain in concrete terms how artificial intelligence is currently transforming the work of music publishers?

[Margaret McGuffin]

Absolutely. AI has been in the studios for a decade. It's used as an ideation tool in studios. Our small companies are using it to make sure they can scale quickly and can compete against global companies.

However, I also want to focus on the fact that in the music industry and in music publishing, we have data, and we are building technology to handle the licensing. What we need is the ability to license by making no changes to the Copyright Act. Data is not the problem in the music industry. We retain our data. However, we want to, in the AI world, make sure that there is transparency of the training data so that we can then license.

[Bienvenu-Olivier Ntumba, Liberal, Mont-Saint-Bruno—L’Acadie, QC]

Thank you, Ms. McGuffin.

Ms. Wilhelm, do you see a model for regulating artificial intelligence elsewhere in the world that Canada could adapt to its music industry?

[Kelly Wilhelm]

I think you're talking about music in particular.

As the others have said, the music industry in Canada is actually very well equipped to do this work. With the existence of SOCAN and with the way in which data has been managed in the music industry up until now, I would say it is one of the industries in Canada that can probably show us the way. Again, we require that clarity in order for that licensing to take place.

There are many other countries that are working in the same way. There's often a comparison to the EU AI Act. It has some clear protections for creators. Again, it comes down to what a country does with that EU AI Act, and that's the level at which we need to be thinking in Canada as well, with the capacity of the sector coming into play, with the clarity of the Copyright Act and then with that investment in creating those marketplaces for the work.

[The Chair, Liberal, Lisa Hepfner]

Thank you.

I'd now like to welcome Alexis Deschênes, who is with us today.

Mr. Deschênes, you have the floor for two and a half minutes.

[Alexis Deschênes, Bloc, Gaspésie—Les Îles-de-la-Madeleine—Listuguj, QC]

Thank you very much, Madam Chair. Thank you for your welcome.

Good afternoon, everyone.

I'd like to address one risk in particular, to see how real and close it is. We read that it's possible to create fictitious groups using artificial intelligence. We recently saw the example of Velvet Sundown, which was able to get a million hits on Spotify.

Is there a real risk that artists and authors will be absolutely left out?

[Patrick Rogers]

The biggest risk to AI slop is stream manipulation, which is music that exists purely to be quickly generated to swamp services and then have people pretend to listen to it or have spins artificially generated. That hurts the overall pool for streaming services. Canada's major labels, as well as the major labels around the world, have joined forces to try to find ways to stop that from happening, because that dilutes the pool and hurts artists everywhere.

[Margaret McGuffin]

The ICMP has done a lot of work with international publishers to look at clones and fakes. They looked at the EU app stores and found many apps that were cloning voices, images and songs. It's pure copyright infringement and the apps need to be taken down. Patrick was talking earlier about having to deal with privacy and security legislation around likeness and image.

[Alexis Deschênes, Bloc, Gaspésie—Les Îles-de-la-Madeleine—Listuguj, QC]

Mr. Ducharme, do you have anything to add?

[Marc-Olivier Ducharme]

I don't have anything to add about music creators specifically, because that's not my specialty. I mainly represent independent artists who do research and development, so that's not exactly our field.

[Kelly Wilhelm]

I would add only that this really speaks to the need for the transparency of labelling what AI-generated content is. We need to know when we have it, when we see it and what the impact is on the business models.

[Jennifer Brown]

I agree. As has been said before, it's the dilution of revenue that is due to human creators right now. By flooding the systems with all of this...and then manipulating it is beyond worse; that's called fraud. You want to make sure you're not diluting the pool and the revenue that's supposed to go to the human creators when their works are performed.

I want to go back to data on this point. We process billions of performances every quarter. Again, there's no fear of what we can handle from AI. The reporting can be handled. We can make sure that licences happen and people properly get paid.

[The Chair, Liberal, Lisa Hepfner]

Thank you.

Mr. Diotte, you have the floor now for five minutes.

[Kerry Diotte, Conservative, Edmonton Griesbach, AB]

Thank you very much.

This is a general question that I think gets to the heart of everything. To each one of you, how much government control would you like to see on AI? What does it look like?

Let's start with Ms. Brown.

[Jennifer Brown]

We've all been pretty clear. First of all, keep the Copyright Act intact, with no exceptions, because when that is clear, it should bring people to the table.

We believe there can be a licensing market. We have market negotiations all the time. Amazon, Spotify and Apple Music have all negotiated contracts with SOCAN, the CMRRA and the labels to pay for the copyrighted work. They're still innovating and they're still paying for the licensing. We believe that can absolutely happen.

The transparency will help us because it will give us the level playing field to make sure that we can absolutely say, “You have our works. This is how we can start licensing.” The clarity that there will be no exceptions will bring them to the table.

[Kelly Wilhelm]

I'll throw in the same points around enforcing the Copyright Act. I think we've been clear about transparency. We're clear about identifying what is AI-generated content, because that matters a lot from the consumer perspective and the perspective of compensation.

It's important to remember that the government has many different roles to play. Legislation is one way it can help. There are direct investments in AI compute in the private sector that are absolutely massive. They could be matched or thought of more broadly in terms of their use in public AI investments, as we've talked about.

Skills and training are a very important part of this to be able to raise the bar for creative industries and to be able to participate in this economy fully. These are often folks who are working for the content and, in the case of independent artists, to get the art on the stage and do the work. They don't often have the ability to invest in that back end. Helping with skills and training development is a really key piece of all of this.

It's similar to anything to do with exports and creating those marketplaces. There are other things the government can do in addition to legislation that are really important to keep in mind.

[Marc-Olivier Ducharme]

I agree with everything that was said earlier.

Last year, less than $1 per Canadian citizen was invested in direct support for artists. In fact, the total amount was about $36 million. Someone asked us earlier why we don't create our own systems. You have the answer. When investments in direct support for artists are $36 million, there's no other way to work.

Independent artists are the ones who support the industry. They're at the bottom of the pyramid, in a way, but they move and flow through it. Video games as a creative industry wouldn't exist without visual artists. Support is needed for that to exist.

Artists are among the lowest earners in Canada. They are in an under-represented job class, and generally don't have the chance to make their voices heard in places like this, and don't hire lobbyists to put pressure on the government.

We're asking the government to invest in technologies, but we want those investments to be controlled by artists.

[Margaret McGuffin]

For the first time in my career, I'm coming here to say, “Do not change the Copyright Act.” It is a low-interference and no-cost solution that will protect Canadian songwriters and the music industry that invests in them.

We know that many AI companies are coming and talking to you and saying how complicated it is. That is not true. They're saying that music on YouTube is free. That is not true. They're saying that it is too many lines of data. That is not true. Our companies are also tech companies. These companies pay engineers and they pay dividends, but they refuse to pay for the songs that they're using to train or for their outputs. That needs to change.

[Patrick Rogers]

I'd like to just agree with everything from Jen and from Margaret specifically.

I'll maybe talk about what it looks like. What it looks like is an understanding by tech companies that they have to get into a licence. What it ends up looking like is licensed legal services that do not yet exist but that we're excited about happening. It looks like a Canada that is a leader in this. At the Kananaskis summit, Canada led the way in saying that we need to do all these great things with AI while respecting IP rights. I think it's important. That was an excellent choice of words in terms of “respect”. It's not “deal with” or “manage” or “figure out”; it's “respect”. I think that's the answer going forward.

[The Chair, Liberal, Lisa Hepfner]

Thanks very much.

Ms. Royer, you have the floor for five minutes.

[Zoe Royer, Liberal, Port Moody—Coquitlam, BC]

Thank you, Madam Chair, esteemed colleagues and, of course, our amazing panel. I'm learning so much from the conversation.

My question is for Ms. Wilhelm.

Prior to being a member of Parliament, I was elected locally in my riding of Port Moody—Coquitlam for 14 years. For the past three years, I was actually a school trustee for the Coquitlam school district, which represents about 4,000 staff and 33,000 students from kindergarten to grade 12.

You mentioned AI skills training and education beginning in K to 12. I know that for one of the larger school districts that might be easier, but there are many smaller districts throughout B.C. and throughout Canada. How do you recommend that implementation? What would be some best practices?

[Kelly Wilhelm]

I'm very gratified to be thought of as having an opinion on this that could be useful. I'm not an expert in education policy, but I do have a husband who is a music teacher.

I grew up and was surrounded by teachers in that capacity as well, so education is very important in my life and, I think, in this case.

When starting at K to 12, the place to begin is with literacy. It's around understanding media, understanding the world around you, knowing what's real and what's not, and having the cognitive skills and the judgment to be able to know the difference. It isn't necessarily that I'm saying that at kindergarten you should be using AI generative tools. That's not necessarily the way forward. However, we do need to have, at the youngest possible age, a skill set for our children to be able to know the difference for what they're looking at.

The challenge is there; I don't disagree with you. There are many sizes of school boards. There are many different approaches to curriculum across this country, depending on where in the country you live. I would suggest that if we're able to do things like the communications technology courses that are mandatory in Ontario now in grade 9, then surely we should be able to figure this out for K to 12.