I cloned myself: The reality of living with a digital twin

Synthesia, a London-based startup recently valued at $4 billion, invited me to their new New York office to see their technology firsthand. Beyond simple scripted videos, the company is pushing into agentic platforms like 'Roleplay Sessions,' which allow employees to practice sales pitches with AI avatars that score their performance. Intrigued, I sat for a two-minute voice recording and a series of photos to create my own digital twin.

My avatar operates on a complex stack—synthesizing voice-to-text, language processing, and video animation models. I tested two versions: a personal avatar that reads scripts and an interactive one trained specifically on my investigative reporting. While the technology is impressive, the results were polarizing. Friends found the interactive model both fascinating and unsettling, particularly when it refused to answer personal questions, instead dutifully redirecting every query back to my research on venture capital fraud.

This experience forces a confrontation with the future of media. While an avatar can handle repetitive tasks or routine briefings, journalism relies on human connection and trust—elements that cannot be coded. Even so, the utility for corporate communication is clear. As I watched my digital double wait silently for input, I couldn't help but feel a tinge of unease. We are moving toward a world where presence is no longer tethered to biology, and while I can simply log off when things get weird, the boundary between the creator and the simulation is disappearing.