If nothing else we're about to be inundated in confabulated texts and AI slop. Now I think this is less of a problem than usually assumed in that there already exists an infinite procession of human generated garbage text on the Internet which gets filtered out by various social grading mechanisms.

On the other hand quantity has a quality all its own and LLMs provide the opportunity for expanded Sybil attacks on the social grading mechanisms themselves. It's simply much easier in 2024 to make up a believable fake person with fake photos, fake social media feeds, and fake opinions.