AI deep-fakes are going to become a major problem in the years ahead - Yorkshire Post Letters

I first became aware of AI deep-fakes on YouTube when watching a talk by Warren Buffet, the renowned investor and multi-billionaire. It looked and sounded exactly like him, but some ten years ago, and what he had to say was quite out of his orbit.

He was talking about Russia taking Odessa and winning the war. A day or so later I watched Professor John Mearsheimer of Chicago University talking about the war in Ukraine - and again about Russia taking Odessa. It was his voice and his image but he appeared too egg-headed and was speaking too quickly to be true.

Both were AI fakes, presumably engineered by actors of the Russian state. Next I watched another YouTube talk, this time an interview on the Un Herd channel of the real former Greek finance minister and now political commentator Yaris Varoufakis.

Here he explained that he had been the victim of AI deep-fakes which appeared so authentic that they had received many thousands of likes and comments. It took several attempts and 10 days for YouTube to remove them, but as soon as one was deleted another popped up.

Mr Varoufakis said that some of the AI fakes were again by Russian actors, but others were there just to gain views. He also had been regularly impersonated on X. Clearly, AI deep-fakes are going to become a major problem in ensuring authenticity in the years ahead.