Sure. There are a number of risks. As you noted, misrepresentation is not new, but the technology now is such that it's very difficult even for experts to differentiate between AI-modified images or created images and actual images. There's a real risk that the technology could be used to have certain important people in the electoral process, like party leaders or the Chief Electoral Officer, be made to say things that they have not said.
There's also the concern, and I speak to that in my report, that people use chatbots to query and to obtain information about all kinds of subjects, including the electoral process. In some cases, that information is incorrect. There are some chatbots, like I believe Microsoft does, that refer electors instead to the proper authority rather than trying to allow their system to create an answer that could be based on incorrect information. In my more recent report, I have made recommendations that platforms who put out AI chatbots refer electors or refer people who have queries to the appropriate authority.