DEEPFAKE FEARS — Canada’s music record labels are pushing the Liberal government to tighten federal law to defend against damage from AI-generated deepfakes.
“We worry that eventual regulation will not come fast enough to prevent some of the more significant harms facing the creative industries and Canadians more broadly,” PATRICK ROGERS, the head of Music Canada,wrote in a letter to the Commons industry committee as it considers Bill C-27, the Liberal government’s digital charter that deals with privacy and AI. The group is seeking to get federal publicity rights spelled out in Bill C-27 to extend protections for Canadians on how their voice is used online, a lobbying push spurred by a fake Drake song. That comes as U.S. lawmakers are making a similar push through the No AI FRAUD Act. Playbook spotted the letter and got on the horn with Rogers. This interview was edited for length and clarity. What's the concern for artists? Historically, the music industry has been a really helpful canary in the coal mine through multiple technological revolutions. The reason music was stolen in the digital download age, before video games and movies, was because the file size was small enough to share over a phone line. In this way, music has been pretty easily and pretty quickly turned into voice clones and deep fakes. We have firsthand experience on this and we wanted to make sure that we got out in front of decision makers who are making these sorts of early framework decisions and C-27 with sort of our biggest concerns of the costs, both economic and social, on what happens when you're when your voice is literally cloned or your image is reproduced in a way to put someone else's words in your mouth. We've had some instances of this already in the music industry with mainstream artists, and increasingly, we're seeing the risks of this to emerging artists. There are antiquated analog laws already on the books, mostly provincially, but some federal things that prevent celebrities from being misused in magazine ads. You can't put George Clooney’s face next to a watch and say he supports your watch without him consenting. We want to make sure those same laws are revamped and retooled and re-understood as being important in the digital space as well. We saw this last year with the deepfake Drake song that had fooled some people as a big example in the media. What's the concern for smaller artists? This is a concern for small artists. It's a concern for everyday Canadians. We're seeing stories of very deepfakes making their way into high school social circles and the damage that can be done there and weaponized against Canada's youth. It's those same issues. There's so much of people in the digital space that their voices and style can be stripped down from these services and made available to people. We want to make sure that when it's clear that something like this has been done to deceive and mischaracterize a person who has not consented to this, it should be fully understood as illegal and then our laws should be updated to make it so. You point to unauthorized deepfakes as a risk to human artists, but also to politicians. Are you hoping to drive the point home by making it more compelling for the lawmakers who are going to read that? The technology that interests you in the fake Drake track is the technology used to cause real civil harm. We do not want to see this technology evolve into a place where it could affect our elections. We don't want to see it in places where people of public prominence seem to be saying things that they never said or never would say. Are you taking cues here from the U.S. bill, the No AI FRAUD Act , that would do the same kind of thing — give people more control over their likeness and digital replicas? It's something we really believe in and so we're happy to see bipartisan progress in the U.S. on this. If they can do it with their First Amendment goals in mind, we're pretty confident we can do it here as well. Have you brought this up with the Heritage Department? Yeah. We've been pounding the pavement now for several months on this. The result has been really positive and really warm. The first response we get is, ‘Isn't this covered by a provincial tort law?’, and that sort of thing. Why we've gotten more aggressive about it is, whether it is or isn't covered by provincial tort law, we think there's a real moment here for federal legislators to stand up and say this should be illegal. The industry minister seems to be in the driver's seat on the amendments for C-27. So if the government doesn't end up addressing this, because they seem to have already come out with their amendments, should there be some kind of separate legislation? Digital regulation, digital laws aren't going to end with C-27. And so if this isn't addressed directly in C-27, we'll look to work with the government to address it elsewhere. — In related reading from POLITICO: Taylor Swift deepfakes nudge EU to get real about AI.
| 
