Ultimately it is just the same problem as with digesting information created by humans. Humans also misinform, out of malice or because they don't know better. The solution is rationality, I suppose. You have to verify the claims that are being made. The bots of the future will just have to get better at that.