The question is not simply whether we can build believable AI avatars.

The question is whether we can build them in a way that is transparent, trustworthy, useful, governed, and worthy of the audience's belief.

Artificial Intelligence is no longer working quietly behind the scenes. It has stepped in front of the camera.

In this demo-heavy keynote, AI pioneer and CTO Tim Huckaby shares the real-world journey of building a real-time, interactive AI avatar for live broadcast television. Combining generative AI, large language models, retrieval-augmented generation (RAG), agentic AI, voice synthesis, computer vision, edge computing, and live production systems, the project pushed beyond traditional AI applications and into an entirely new category of audience experience.

As AI avatars become increasingly lifelike, audiences naturally begin to assign human characteristics, intentions, and credibility to them. This phenomenon, known as anthropomorphism, creates tremendous opportunities for engagement, education, entertainment, and accessibility. It also raises important questions about trust, authenticity, disclosure, consent, accountability, and the growing responsibility of organizations deploying synthetic personalities in public-facing experiences.

Drawing on lessons learned from taking an AI avatar from concept to production, Tim shares technical successes, architectural failures, unexpected surprises, and practical insights into delivering low-latency, real-time AI interactions. He also explores the role of open and interoperable technologies that enable AI systems to operate across cloud, edge, GPU, and media production environments where reliability and responsiveness are critical.

Through live demonstrations and real-world examples, attendees will gain a firsthand look at the future of AI-driven media and communication, and the governance challenges that emerge when AI evolves from a tool into a personality.