This looks like an attempt to build good will with the governments of the world so that their platforms don't get crushed by regulation in the coming years. And also a way to give their supporters ammunition to fight a total AI ban by noting all the good and necessary uses of AI. The accord points out several times, for example, how AI technology can be used not only to generate deep fakes, but also to detect them.
Considering just how effective their claimed attempts at stopping the spread of ordinary fraud, scams, fake reports and so forth have been up to now, color me doubtful as fuck about their pulling it off with AI deepfakes. If anything, these very companies will be what most helps the spread of this kind of content, for elections and so many other things.<p>In brief, this pledge smells of heavy duty PR bullshit for the sake of feel-good political goodwill.