In April of this year, the boss addressed important words to his employees. Jensen Huang, CEO of the tech group Nvidia, used teleconferencing software for this, as far as the new normal in pandemic times. What the employees did not know, however, was that it was not the real boss who was speaking to them, but a computer-generated image, powered by the high-performance hardware of their own company. None of the employees, among whom Huang has, mind you, a status quasi-redeemer, noticed the swap.
Huang’s appearance was more than just a marketing gimmick. Images, videos and sound snippets generated from nothing by artificial intelligence, so-called deep fakes, were only a theoretical phenomenon just a few years ago. Technology critics warned of their corrosive potential: What would happen if every politician could be artificially staged in front of any background and with any O-tones? When synthetic humans become key witnesses for every conspiracy theory imaginable? Wouldn’t that be disastrous for the already heavily battered public discourse? And finally, what about banal, completely agendaless revenge? With the help of technology one could credibly mount the ex’s face in porn movies and thus ruin the reputation of every person in the long term.
As the example mentioned above shows, deep fakes are now good enough to deceive even people who are extremely familiar with the fake person. The world has once again turned too fast to establish new, contemporary rules. In addition to undermining democracy, the artificial humans with computer-generated attributes turn a whole range of other traditional systems upside down. The guild of voice actors, for example, is facing dark times. Because voices generated by AI are now so lifelike and authentic that professional speakers are hardly needed anymore.
You can sell your face to an AI company, which then plays copywriting on the person
Capitalism would not be itself if it did not ensure that people even try to be exploited. Startups like Israel-based Hour One pay people for the rights to their faces. All you have to do is take a seat in front of a high-resolution camera. If you feed the resulting data into an AI software, the company can use it to generate an endless amount of video material in which the person in question says what they want in any language. The clips generated in this way are to be used in promotional videos for companies that need credible testimonials. Customers select a face, upload the text for it to say, and receive a video that looks like a real person is speaking the text into a camera. Every time their own face becomes part of a new campaign, the donors receive a small royalty payment. It goes without saying that you have nothing to say when it comes to which message your face will be used for in the future.
In view of the current developments, an almost comfortable predecessor technology is stock photos. For example, you could secure a small additional income during your studies by having your face photographed in scenarios that are as generic as possible, which are then found by photo editors in advertising agencies using keywords such as “happy family” or “dynamic team”. You just had to say goodbye to sovereignty over your own face. The greatest danger was that one would find one’s former self at some point as part of a CDU image campaign.