Tel Aviv-based startup Hour One is paying people to use their likenesses to create AI-voiced characters for marketing and educational videos for its clients
AI-powered characters based on real people can star in thousands of videos and say anything, in any language. Tweets: @strwbilly , @timsowula , @pomeranian99 , @carolhtucker , @giano , @funnymonkey , @combsthepoet , @iwillleavenow , @farahshoucair , @i_m_shweta , @domwitczak , @hackermaderas , @mat , and @hypervisible Tweets: Will Douglas Heaven / @strwbilly : People are hiring out their faces to become deepfake-style marketing clones: “I can't say that anyone today is making a living doing this. But we think if all goes well it will be a viable way to make an income.” https://www.technologyreview.com/ ... Tim Sowula / @timsowula : I may be a bit of a luddite but I can't see this ending well: People are hiring out their faces to become deepfake-style marketing clones https://www.technologyreview.com/ ... Clive Thompson / @pomeranian99 : An AI firm is paying people to use their faces in deepfake-ish corporate videos — i.e. they scan your face while you make expressions and talk, then they can autogenerate you saying *anything* for ads, promo, customer service, etc: https://www.technologyreview.com/ ... Carol H Tucker / @carolhtucker : While I have always been considered a bit of a character in RL and my 2nd lives, I'm not sure I'm ready to let myself become part of the deep fake industry.... https://www.technologyreview.com/ ... @giano : This is a must-read article about applied #AI regardless of the sector you work in (but especially important if you are in marketing). How will we sell ourselves in the future? > People are hiring out their faces to become deepfake-style marketing clones https://www.technologyreview.com/ ... https://twitter.com/... Bill Fitzgerald / @funnymonkey : Another day, another tech startup finding ways to extract the humanity from people. https://twitter.com/... Tawana Petty / @combsthepoet : 🙄😬 “The firm does not let people have a say in how their likeness will be used or what words will be put into their mouths, but it has an ethics policy... “We're pretty conservative about the types of businesses that we work with,” ... no gambling, no sex, and no politics.” https://twitter.com/... Techni-Calli / @iwillleavenow : Can't imagine this causing serious reputational harm or confusion for anyone, nope, no siree. https://twitter.com/... Farah Shoucair / @farahshoucair : Hour One uses a camera to film a person talking and making different facial expressions. It plugs the resulting data into #AI software to deepfake tech, generating an endless amt of footage of that person saying whatever it wants, in any language. https://www.technologyreview.com/ ... Shweta / @i_m_shweta : Such a fine line between what's ethical and what's not. Wonder how long will it take it to be used beyond marketing and educational videos. #AI #aiethics https://www.technologyreview.com/ ... Dominik Witczak / @domwitczak : https://www.technologyreview.com/ ... Makes me wonder if anyone from these people ever stopped & wondered for a second, what would've happened, had the pre-trained AI model data ever leaked? @hackermaderas : #CyberpunkisNow Companies are ‘hiring’ peoples physical appearances for use in commercials created via deepfakes-like software. The resulting AI-powered characters (based on real people) can star in thousands of videos & say anything in any language. https://www.technologyreview.com/ ... https://twitter.com/... Mat Honan / @mat : “A human being doesn't need to waste their time filming.” Deepfake acting is heading your way https://www.technologyreview.com/ ... Hypervisible Dot Pdf / @hypervisible : Startup pays people for the rights to their face and makes deepfake “characters” that businesses can hire for things like appearing in videos and ads. The person receives a micropayment anytime their face is used. https://www.technologyreview.com/ ...
Context & Ripple Effects
Hour One is extending a line of work that started with Synthesia's corporate-friendly deepfakes for multilingual training videos and picked up momentum just weeks earlier when startups like WellSaid Labs and VocaliD began building custom AI voice actors. What is new here is the supply side: instead of synthetic avatars, Hour One recruits real people and pays micropayments for ongoing rights to their faces and voices.
The company frames this as nascent income — its own talent lead admits nobody makes a living from it yet — while critics in the coverage flag the ethical and reputational risks of renting out a likeness that can then say anything, in any language, on behalf of paying clients.
First-order effects
- Hour One's enrolled presenters exchange broad, perpetual usage of their likeness for micropayments, while client brands gain characters that can star in thousands of marketing and educational videos without the human being involved after sign-up.
- Critics quoted around the story immediately frame the model as deepfake-adjacent, putting reputational scrutiny on Hour One and any brand that deploys its characters.
Second-order effects
- Rivals in the corporate-video space — most directly Synthesia, whose multilingual training-avatar business established the category — face pressure to match a real-human-likeness supply chain rather than rely on fully synthetic presenters.
- The voice-actor licensing model built by WellSaid Labs and VocaliD gains a visual counterpart, pushing agencies and video-production buyers toward a market where a face-plus-voice license replaces per-shoot talent fees.
Third-order effects
- If paid-consent licensing becomes standard, likeness governance — scope, duration, and revocability of face and voice rights — turns into a contract and regulation battleground, especially once the same tools are shown enabling abuse: a later WSJ test of an AI clone fooling a bank's voice biometrics demonstrated the fraud vector, and Microsoft's VASA-1 portrait-to-video model shows the technical bar falling to a single photo.
- The longer-run structure splits synthetic media into two tiers — consented, compensated likenesses versus unconsented clones — forcing platforms and clients to prove provenance, which favors licensed-talent platforms like Hour One only if verification becomes a buyer requirement.
The trend: Human faces and voices are becoming licensable raw material for synthetic video, with consent-based platforms racing ahead of both regulators and models that require no consent at all.