TED imagines its 40-year future using OpenAI's Sora

What does a TED Talk look like in 40 years? For TED2024, the conference decided to find out using generative AI.

TED partnered with director and artist Paul Trillo and OpenAI to create an exclusive concept video, built entirely with OpenAI's unreleased text-to-video model, Sora.

Reimagining the stage with Sora

The visual short imagines where the iconic ideas conference could go over the next four decades. It is the latest example of OpenAI keeping Sora in private preview, handing the keys to a select circle of prominent artists and filmmakers.

While the model remains locked behind closed doors for the public, these targeted creator showcases reveal how Sora handles complex, surreal visual worldbuilding when guided by an experienced director.

Why it matters

For creators and filmmakers, collaborations like this provide the clearest look at where text-to-video technology fits into professional production pipelines. You still cannot use Sora yourself, but it is increasingly clear that top-tier creatives are already using it for conceptual filmmaking.

We think OpenAI’s drip-feed strategy makes a lot of sense. By restricting access to high-profile creative partners, the company builds cultural prestige around Sora while buying necessary time to tackle compute demands and safety guardrails before a broader rollout.