AI text watermarking and quality loss

Automated excerpt

But some argued that watermarking does not affect quality at all: it’s essentially just using different random numbers to generate the text. They go on to explain: when sampling several responses to the same prompt, the similarity between the responses is greater for the watermarked responses than the unwatermarked responses. Well, now you’re affected by watermarking, because you have hit the “inter-response diversity” problem: like throwing a loaded die, you’ll see similar responses more than you would with an unwatermarked model.

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.