Synthesia builds first journalist avatar, trained on one story

Synthesia, originally based in the U.K., has grown alongside other digital avatar startups like D-ID, HeyGen, and Colossyan, reaching a $4 billion valuation and crossing $100 million in ARR. Its products include classic scripted video avatars, an agentic platform called Sessions for surveys or roleplay, and an API platform that lets customers combine Synthesia's video and voice models with other tech services. For the reporter's avatar, the team used a mix of voice-to-text, language, text-to-voice, and video models, with Synthesia's own video and voice models and options from other labs. The avatar was made in a couple of days after a photo session and a two-minute voice recording.

The experience raised questions about journalism and trust. The reporter noted that the biggest part of journalism is trust, which doesn't seem like it could ever be outsourced to an AI. One investor immediately said no to avatars presenting the news, while others weren't so sure. Friends found the avatar both interesting and creepy, and her mother joked about giving birth to two of her.

Outside journalism, the idea of cloning yourself could be appealing for handling work questions. The reporter came away with mixed feelings, noting that deterministic avatars will never blink or say something new. Whether avatar use spreads in corporate America likely hinges on acceptance and trust, especially for younger generations.