It's a bit funny they've made this about Musk. He typically seems to underplay risks of tech, but I suspect that is his genuine view, he is one of those people who puts altogether too much trust in the idea of technological progress. It's a kind of subset of progressivism as a political ideology.
But the issues with AI really aren't about Musk, if he fell off a cliff tomorrow nothing would change. AI would still steam ahead.
My guess is that the ability of AI to create images like this will in a certain way decrease their power. SO someone makes an AI image of me licking chocolate off of Elon Musk, or something? So what - it's AI. Everyone will know it's AI, and even where the image is plausible, it will create massive deniability.
Of course this will create all kind of it's own problems, which are an extension of the trust and credibility issues the information age has unleashed. Shorter term I predict some big money in certifications to "prove" what is real.
The only think I suspect may slow this down is that while AI is cheap or even free now, once we are dependent on it we will surely have to pay for it, more and more. That's been the model for some time, introduce new technological applications for low or no cost, make people dependent, then jack up the price incrementally over a few years.