[jka]

A brief thought experiment, followed by a suggestion about how to combat the worst of this effect, in the unlikely case that it became true:

Even if all of the content you're interacting with is managed and generated, it does still ultimately affect the real world around you in some way. Call it "fake" if you will, but it has a downstream effect. You can ask your friends and neighbours whether they have the same opinions about the world, whether they've read the same articles and reached similar (or opposite) conslusions, and see the results of political and regional decision-making around you.

The strategy is to keep a clear vision of what you think improvement is, and to require proof-of-work from others during interactions that bring you (ideally all, in a multi-person setting) closer to your vision.

If an entity that you're interacting with (whether it's a person, an organization, a bot, or entirely indeterminate) doesn't provide verifiable progress (ideally without loopholes) towards your vision, that's when you need to decide whether to attempt to correct their actions (perhaps aided by the likeminded group around you), or simply to gradually disengage if they don't appear to adapt and realign over time.

You should also be aware that you and your own vision may be realigned by the proof-of-work requested by others; and in those cases you also have a value judgement to make: do I change my opinions based on the requests and evidence I've received, or do I continue to adhere to my chosen path?