It's their report, their claims, their argument. The burden to make the case and be transparent is theirs, not mine. It's dishonest to present AI summaries as facts.
There should be at least a sample of the source text so we can see how well the descriptions match the sources.
Also, where is the prompt, the model? The methods used to present the content.
It's funny how much critique of "AI slop" is on HN, and here's a case where even mild hallucination is very misleading, yet the bar is incredibly low.