I created an interactive digital avatar of myself — and you can talk to it

It emphasizes transparency in synthetic media use and warns against misleading

Alexandru Voica, head of corporate affairs at Synthesia, shared a link this summer to an interactive virtual version of himself designed to handle media inquiries. The avatar was trained to answer common questions about the company’s video-generation technology. This followed a panel discussion where Voica had spoken about AI ethics and synthetic media. The avatar represents Synthesia’s effort to scale communications using its own generative AI tools. It was built using the company’s platform, which creates lifelike digital humans from text input. Voica explained that the goal was to test how well an AI-driven representative could manage routine press interactions. The system draws from pre-approved responses and adapts to variations in questioning. How the avatar handles venture fraud discussions When prompted about venture fraud in the AI sector, the avatar delivers responses based on Voica’s prior statements and company positioning.

Breaking news

Trump Orders Federal Agencies to Adopt „Super Intelligence” TerminologyThe Silicon Valley Accountability GapRunning a Local AI Model on a Phone Handles Most Chat PromptsGoogle Photos Could Soon Offer a Fresh Start with Ask Photos Feature

It emphasizes transparency in synthetic media use and warns against misleading applications of deepfake technology. The avatar avoids speculation and sticks to verified talking points about Synthesia’s safeguards. Voica noted that while the avatar performs well on factual queries, it struggles with nuanced or hypothetical scenarios. Can an AI avatar truly represent a person’s judgment? Voica admits the avatar lacks true understanding and cannot replace human judgment in complex situations. He views it as a tool for efficiency, not a substitute for authentic dialogue. The experiment revealed limitations in contextual Still, he sees value in using avatars for repetitive outreach, freeing humans for higher-level engagement. The project remains internal, with no plans to deploy it publicly without further refinement. Frequently Asked Questions How accurate is the avatar in answering press questions? It handles routine inquiries effectively but may oversimplify or miss subtleties in complex topics.

Can the avatar detect sarcasm or irony in questions? No, it processes language literally and does not interpret tone or intent beyond keyword recognition. Is the avatar available for public use? No, it is currently an internal prototype used only for testing and development purposes.