Deep fakes are ‘a scary phenomenon’
ARTIFICIAL intelligence technology which can be used to make sophisticated fake videos that “threaten democracy” is being funded by taxpayers.
“Deep fakes” is a term used to describe artificial intelligence that mimics facial expressions. It can be used to build propaganda videos by making anyone seem to say things they have not said.
Synthesia, a British company, is a pioneer of the technology. In an era of fake news, senior politicians in Britain and the US have expressed alarm at the threat posed by deep fakes to spread misinformation or stir up conflict.
Damian Collins, who chairs a parliamentary select committee on fake news and has been a vocal critic of Facebook and its response to the Cambridge Analytica scandal, said deep fakes were “one of the most serious threats that democracy faces”.
Marco Rubio, a US senator, has declared the technology a propaganda weapon, and recently several American politicians said it was a national security threat. Despite this, Synthesia, based in north London, received £69,384 from Innovate UK, the government-funded research and development body, on Jan 26 this year. The firm states the goal was “to construct a low-cost, class-leading, three-dimensional (3D) morphable model of the face that focuses on representing the detailed dynamics of the mouth in speech to achieve unprecedented levels of realism in video synthesis of human speech”. Victor Riparbelli, chief executive of Synthesia, said it was “super-aware” of the threat posed by its technology, but claimed it was being developed for use in media and entertainment for dubbing.
“We are not releasing any software to the public… There is a wider discussion to be had about the malevolent use of this kind of stuff,” he said.
One of the biggest investors in Synthesia is Mark Cuban, a billionaire US technology investor who has presidential ambitions. Documents seen by The Sunday Telegraph show Mr Cuban’s investment vehicle, Radical Investments, has invested $1m (£780,000).
Ben Nimmo, an expert in online propaganda at The Atlantic Institute, said: “Deep fakes are a deeply scary phenomenon.” Sheryl Sandberg, Facebook chief operating officer, told US senators recently that “we are working on detecting deep fakes”. Mr Collins told The Telegraph the technology presents a “clear and present danger” by creating opportunities for “black ops” against politicians during elections to destroy their reputations.
A spokesman for Innovate UK said: “Innovate UK supported a short-term feasibility study into how artificial intel- ligence could be used within the specialist film and television industry to improve the quality of dubbing into different languages.
“As part of their application the company identify the potential risk to public trust through the broader application of their technology and provided reassurance that they are working with partners to manage that risk in future.”
Robert Chesney, a law professor at the University of Texas, wrote earlier this year: “The spread of deep fakes will threaten to erode the trust necessary for democracy to function effectively.” Mr Riparbelli proposed that in future videos could contain a watermark to show they have been edited.