It's the difference between drowning in an ocean of bullshit and living on a planet that is one giant bullshit ocean.
AI scales much better, than human bs generators. I consider at least this aspect concerning.
There are 7 billion humans on the planet now, we don't have a problem generating bullshit at scale. The same methods you'd use to discern AI generated bullshit are already needed to evaluate current bullshit. If anything I'd argue that this kind of tech becoming common place could get people to start evaluating all news more critically.