Madam Chair, committee members, thank you for the opportunity to speak before you this afternoon.
By way of brief background, my name is Keita Szemok-Uto. I'm from Vancouver. I was just called to the bar last month. I've been practising, primarily in family law, with also a mix of privacy and workplace law. I attended law school at Dalhousie in Halifax, and while there I took a privacy law course. I chose to write my term paper on the concept of deepfake videos, which we've been discussing today. I was interested in the way that a person could create a deepfake video, specifically a sexual or pornographic one, and how that could violate a person's privacy rights, and in writing that paper I discovered the clear gendered dynamic to the creation and dissemination of these kinds of deepfake videos.
As a case in point, around January this year somebody online made and publicly distributed sexually explicit AI deepfake images of Taylor Swift. They were quickly shared on Twitter, repeatedly viewed—I think one photo was seen as many as 50 million times. In an Associated Press article, a professor at George Washington University in the United States referenced women as “canaries in the coal mine” when it comes to the abuse of artificial intelligence. She is quoted, “It's not just going to be the 14-year-old girl or Taylor Swift. It's going to be politicians. It's going to be world leaders. It's going to be elections.”
Even back before this, in April 2022 it was striking to see the capacity for, essentially, anybody to take photos of somebody's social media, turn them into deepfakes and distribute them widely without, really, any regulation. Again, the targets of these deepfakes, while they can be celebrities or world leaders, oftentimes are people without the kinds of finances or protections of a well-known celebrity. Worst of all, I think, and in writing this paper, I discovered there is really no adequate system of law yet that protects victims from this kind of privacy invasion. I think that's something that really is only now being addressed somewhat with the online harms bill.
I did look at the Criminal Code, section 162, which prohibits the publication, distribution or sale of an intimate image, but the definition of “intimate image” in that section is a video or photo in which a person is nude and the person had a reasonable expectation of privacy when it was made or when the offence was committed. Again, I think the “reasonable expectation of privacy” element will come up a lot in legal conversations about deepfakes. When you take somebody's social media photo, which is taken and posted publicly, it's questionable whether they had a reasonable expectation of privacy when it was taken.
In the paper, I looked at a variety of torts. I thought that if the criminal law can't protect victims, perhaps there is a private course of action in which victims can sue and perhaps get damages or whatnot. I looked at public disclosure of private facts, intrusion upon seclusion and other torts as well, and I just didn't find anything really satisfied the circumstances of a pornographic deepfake scenario—again with the focus of reasonable expectation of privacy not really fitting the bill.
As I understand today, there have been recent proposals for legislation and legislation that are come into force. In British Columbia there's the Intimate Images Protection Act. That was from March 2023. The definition of “intimate image” in that act means a visual recording or visual simultaneous representation of an individual, whether or not they're identifiable and whether or not the image has been altered in any way, in which they're engaging in a sexual act.
The broadening of the definition of “intimate image”, as not just an image of someone who is engaged in a sexual act when the photo is taken but altered to make that representation, seems to be covered in the Intimate Images Protection Act. The drawback of that act is that, while it does provide a private right of action, the damages are limited to $5,000, which seems negligible in the grand scheme of things.
I suppose we'll talk more about Bill C-63 in this discussion, and I do think that it goes in the right direction in some regard. It does put a duty on operators to police and regulate what kind of material is online. Another benefit is that it expands the definitions, again, of the kinds of material that should be taken down.
That act, once passed, will require the operator to take down material that sexually victimizes a child or revictimizes a survivor—