HeyGen

Harvard bootcamp offers AI instructor avatars

Harvard Business School’s HBS Foundry eight-week bootcamp (priced at $699) has integrated AI-generated avatars of its instructors to provide individualized feedback during practice pitches and board meetings. The avatars were created by startup HeyGen and supplement weekly live instructor sessions. New York Times reporter Sarah Kessler tested the system, pitching to an AI-generated version of Flybridge Capital co-founder Jeff Bussgang; participants reported liking the avatars despite some describing them as "creepy." Project director Katharina Rings said the AI component was originally imagined as a chatbot but was expanded after student feedback requested a more guided experience.

Fish Audio raises $52M seed for AI voice models

Fish Audio, a Palo Alto startup building AI voice models, raised $52 million in a seed round led by Coreline Ventures and Capital Today. Since launching last year, the company says it has over 8 million users, $21 million in annual recurring revenue, and a library of more than 15,000 natural language controls. Fish Audio open-sourced several speech-generation models while offering its S2.1 Pro model via paid API and provides creator and enterprise plans. The company faced allegations that some voices were uploaded without consent and has automated its takedown process to remove contested voices within three minutes. Fish Audio plans to release an audio-understanding model and is building a speech-to-speech model.

Google Vids Adds Personal AI Avatars

Google announced an update to Google Vids (part of Google Workspace) that lets users create custom digital avatars that look and sound like them from a selfie and voice recording. The update brings Google’s multimodal AI model Gemini Omni into Vids, enabling video generation from written prompts plus reference images, background swapping, lighting fixes and step-by-step edits. Google says personal avatars will be tied to the account holder’s likeness and Google account and will be invisibly watermarked with SynthID. Access to personal avatars will be limited to certain regions and to users aged 18 or older. The changes shift Vids from an AI-assisted presentation tool toward a broader video-creation platform that competes with AI video startups.