Fake it as you make it? Leading figures at ad industry trade body the IPA have starred in ‘deepfake’ videos to illustrate the opportunities and dangers in AI tech on the industry,
Videos played at a recent IPA Council debate about whether AI is a threat to the advertising business, showed a deepfake version of the former IPA President Julian Douglas and a deepfake Paul Bainsfair – the IPA’s current Director General – taking opposing sides of the argument.
The videos were created by Morten Legarth, creative director at ad agency VCCP, who used AI to produce various aspects of the films. Scripts for the debate speeches were written by GPT 3.5, while Douglas’ and Bainsfair’s voices were generated using Eleven Labs. Face animations for the film were done by D-ID.
Arguing for the motion that AI poses a threat to the advertising business, Deepfake Douglas said: “We must adapt to the changing landscape of our profession or risk becoming irrelevant. I urge all of you to take a closer look at how AI can transform your advertising campaigns and start experimenting with this technology today.”
Deepfake Bainsfair replied that “while AI technology certainly has its benefits, it is not a threat to the future of advertising agencies. Human beings are the ones that build the relationships, develop strategies, and create content that resonates with the intended audience.”
After other (real) Council members made their contributions to the discussion, the side against the motion won the day by a small margin.
Commenting on the outcome of the unusual debate, IPA said: “It appears that until AI can understand nuance, ethics, and the appropriateness of moment – and it no longer requires a talented human being to create these things, even with all the AI tools – then humans still have the advantage. But only just.”