With AI we're about to have a trust crisis (I'll explain in a sec), so if we extrapolate it very far, we could see a world where nobody trusts anything displayed on a device, which is essentially what you describe but in a world with electricity still running.

The trust issue: we already see hacker news members accusing each other of being LLMs, we already see legitimate images being immediately put in the AI category, and AI images and videos being shared as if the truth. We already see signs of infosec changing forever, with flaws found in SSL libraries that makes you wonder how trusty that little padlock next to the URL will become, we already see operating systems having to patch faster than they can.

With all that you get to wonder how much you can trust the pixels you see on your device, and how much we'll collectively trust them in the future. That leads you to trust your neighbours way more than your social media feed, which frankly, may have a positive impact on society over the long run, despite a very painful transition.

That's under the hypothesis that AI leads us to experience a trust crisis, which isn't a given.