Pew recently analyzed the content descriptions of all the leading news podcasts to determine the typical content of the most popular podcasts. The results are interesting, but the methodology was more so. They created a very detailed prompt to get Chat GPT to analyze it. They then tested the results against humans and found the results equally accurate.
The key, of course is to create the perfect prompt that will get you the actual results you are looking for.
Look for a lot of human coding to be replaced by AI in the next few years..