According to TechCrunch, a reporter agreed to let Synthesia create a digital twin of her after visiting the company's new New York office. Synthesia, a U.K.-based startup focused on AI video avatars, built both a personal avatar that reads scripts and an interactive avatar trained only on her article about venture-backed startup fraud. The company says it has crossed $100 million in annual recurring revenue and reached a $4 billion valuation earlier this year.
Synthesia's product lineup now spans a video-creation platform, an agentic platform called Sessions, and an API platform. Its recently launched Roleplay Sessions product lets employees rehearse sales pitches with an interactive AI avatar that responds and scores their performance. The interactive avatars combine voice-to-text, language, text-to-voice, and video models, with options to use voice and language models from providers such as Cartesia, ElevenLabs, Google, or OpenAI.
The reporter found her personal avatar's voice fairly accurate, while friends found the results both interesting and creepy. The interactive avatar refused to answer off-topic questions and kept redirecting users back to the article it was trained on. The experience raised questions for her about whether avatars could augment or replace journalists, though she argues that the trust at the core of journalism is unlikely to be outsourced to AI.