I’m a broadcast engineer (someone who designs/maintains/etc. transmitters and studios for radio stations) and quite new to AI.
Recently I decided to ‘get my toes wet’, so to speak, by running OpenAI Whisper locally and experimenting with transcribing radio show recordings. It’s worked out pretty well.
Now I’d like to take another step, and see about automatically summarizing these transcripts. The goal would be to provide the full text (a typical show might consist of 500 lines of transcribed dialogue), and get back a title (five words or less) and summary (fifty words or less).
If this were to work out, it could open the door to offering archived recordings as podcasts - something I’ve so far avoided due to the labor involved in generating these titles and descriptions.
Can anyone recommend some thing(s) I might look at, to get an idea of where to go from here?