In about October last year, there was some excitement about a new podcast being hosted by an AI re-creation of Michael Parkinson. It got all of the excitement and attention you’d expect at the time, and then – well – it all went a bit quiet. So the other day, when I had an idle moment I found myself Googling it – and I was surprised to discover that it had launched.
The voice is uncanny. The interview less so.
Now other than awe at the incredible synthetic voice and just how incredibly realistic it is (but then, I tell myself, there was a hell of a lot of material for it to go on so of course it is), the podcast is a little odd. The questions feel formulaic, the pace of the interview really consistent (and quite sedentary) and I’m not sure it really feels up to being associated with that great, prolific and talented interviewer.
Where the conversation falls apart
Chris McCausland, the interviewee from the second episode, remarked at the end of his interview how unusual it was for the tone of his answers not to play a part in how the conversation was going. And he was right. Clumsy links between an answer and the next question, repetitive questions on the same topic. It would be unfair to equate the interview to an early-career interview by an aspiring journalist but the lack of tangible connection the back and forth generates certainly makes me feel like being unfair.
What (just about) makes it work
What saves episode two is that Chris gives some fantastic answers, adapts to the pace, and is really self-reflective so – I guess – it still achieves the aim of an interview. And the creators are honest that this is a work in progress. Each episode is followed with a sort-of-behind-the-scenes follow up where how the AI performed is discussed.
A shift in mindset
And I guess having listened to that, I switched my mindset just a little.
This isn’t about a podcast. It’s about funding.
This is clearly an amazing technology that they’re developing and even though I’m not sure exactly of the purpose – a field such as screening interviews for jobs seems more likely to be a good application for this than a podcast, or a TV show – it really did need to be Parky, even though it didn’t.
Not everyone is comfortable with that
Without the name to hook it onto, how could the people behind this technology have secured the funding to develop it? Whether it’s continues to improve or not, the novelty factor of being interviewed by AI (or listening to an AI interview) will give them at least 2 series or so and that means plenty of opportunity to generate the cash needed to pour into the ‘project’ behind it. Even if some people hate the idea.
And then there’s money
Because, of course, Michael Parkinson’s son has found another way to monetise the back catalogue of interviews he owns through this podcast. So it’s wins for everyone. But make no mistake (gosh, I hate that phrase), this isn’t a podcast that revives the much-missed Parkinson interview.
