I’d imagine that you could even take podcasts and run them through a speech recognition app much like visual voicemail does. This could then parse the text and flag a podcast for manual review by a human to ban an account, could even auto suspend the account until its challenged or reviewed. You don’t even need someone to listen to everything since Podcasts and usually spoke word.
Hell I bet I could build a pipeline to run on a local server in under a week that does this. Download the audio. Parse it into text. Then parse the text for any trigger words or phrases.
I’d imagine that you could even take podcasts and run them through a speech recognition app much like visual voicemail does. This could then parse the text and flag a podcast for manual review by a human to ban an account, could even auto suspend the account until its challenged or reviewed. You don’t even need someone to listen to everything since Podcasts and usually spoke word.
Hell I bet I could build a pipeline to run on a local server in under a week that does this. Download the audio. Parse it into text. Then parse the text for any trigger words or phrases.