AI am Scarlett! No, AI am Scarlett!
Is the Centre for Media Transition as interesting as a flirty chatbot with a voice stolen from Scarlett Johansson? Judge for yourself next week, when the CMT will be front and centre at the humAIn conference for media and marketing executives in Sydney on Tuesday 28 May.
Two CMT folks will be on-stage for the panel session ‘AI Read the News Today, Oh Boy’ (and props to the organisers for that title): Michael Davis, who co-authored the Gen AI and Journalism report, and Shaun Davies (that’s me), consultant on AI and media and part-time grad student working with the CMT.
We’ll cover the implications of generative AI in the media. Michael will discuss some juicy moments from the report and I’m keen to talk about how newsrooms can keep themselves safe with human-in-the-loop workflows and robust safety metrics for any AI products they deploy. We’ll be on-stage with Austereo’s Melanie Withnall, who’ll bring her coal-face perspective on working with AI in an actual newsroom, and Future Media author Ricky Sutton, a prominent advocate for newsrooms getting paid fairly when their data is used to train AI.
Scarlett Johansson probably won’t be there, but she may come up in conversation. This week Johansson revealed that OpenAI’s Sam Altman ardently pursued her for permission to use her voice in ChatGPT 4o. Presumably he was inspired by Her. When Johannson turned Altman down, Open AI apparently hired a soundalike to replicate her voice. You can read my reflections on the implications for copyright here.
The big question: should the vibe of an artist’s performance – their tics and voice modulations and mannerisms - be protected under copyright law?
Shaun Davies, CMT researcher