Travis suffered a stroke a decade ago and is largely unable to sing now. Where That Came From was originally sung by James Dupré (he is credited as the “vocal bed” on the single) with AI technology used to fuse Travis’s sound (which was collected over a 15 year span of his career) over Dupré’s.
I’m also still mulling over Sam, the AI-powered voice that Mamamia has trained up on hundreds of hours of the media group’s podcasts. Sam means that Mamamia’s stable of 50+ podcasts can provide for ad clients host voiced (but AI generated) reads. It’s clear from the comments to this story on our Facebook page that Sam struck a chord with you too.
How does this now work for live reads? Do you still get paid per commercial, or, if you have allowed Sam, or something like Sam, access to many samples of your voice, is it a part of your contract that this just happens for clients who want ‘you’?
I don’t think I’m revealing any great industry secret in saying that most radio hosts would prerecord their live reads before their show. Part of the benefit for the client, even with a pre-record, is the adlibbed ‘good will’ that often is done by the client’s chosen talent above and beyond the script. With AI you will never get those sort of freebie extras again. You will get it read verbatim, spelling errors, funky pronunciations and all.
AI is getting better and better, so I do understand the excitement of Mamamia to be at the forefront of the ways this technology can save time.
For me though I see it as akin to the roll out of self serve check-outs in supermarkets. Good staff will no longer have jobs. Theft will increase. Expensive product will be mis-scanned at a cheaper price and the machine won’t be able to identify a feijoa.
What are your thoughts?
