I'm sorry. I thought you were directing that to the Equal Voice folks—my apologies.
There are a few main things. The first piece of it has to do with what's being covered by Bill C-16. Right now, we're covering only the distribution of deepfakes, but we should also be looking at criminalizing the creation and possession of deepfakes, as a preventative measure against further distribution. A couple of ways could be to specify this to be creation and possession with intent to distribute for cases of threats or blackmail, or negligent or wilful possession after consent has been withdrawn.
Just yesterday, the Government of Alberta announced their own deepfake legislation. That includes provisions around creation and revocation of consent, so there's very much a precedent for that kind of approach in Canada, as well as internationally. South Korea has built a really robust legislative framework around deepfakes.
I think we should also be expanding the scope of imagery covered by the law to cover non-consensually sexualized uses of one's likeness, rather than the definition set out in Bill C-16, which covers only imagery that is “likely to be mistaken for a visual recording of that person.” As people in the previous session talked about, there are many ways that this kind of definition can shift the legal burden onto victims and create a lot of lack of clarity in how the law is interpreted.
With regard to Bill C-34, we can also amend that to require providers of generative AI services to demonstrate in some way that their systems have been tested and safeguarded against hateful and non-consensual sexual uses prior to deployment, rather than the current status quo in Bill C-34, which is to create a digital safety plan, but it's not clear if the provider is going to be following that.
