Sora 2 and the Day Hollywood Went to War
Forget Tilly Norwood. OpenAI’s new cinematic model can fake anyone and anything — and power players finally draw their line
This is the second in my two-part series for paid subscribers on OpenAI’s new tools and what they mean for Hollywood. Earlier, I wrote about how ChatGPT-5 has changed writers rooms overnight.
Hollywood finally has hit its breaking point with AI — and the name of that breaking point is Sora. As I write this, I realize that I’ve already covered OpenAI this week in a previous column…
…and that I’ve already used the word “existential” in describing the conflict this company presents to the entertainment industry. And here today, to paraphrase Animal House, I must write about “double secret” existentialism, as Sora 2 could be that much more impactful and dangerous to Hollywoodland.
For a year, the industry has been pretending it could live with AI enhancements, virtual assistants and text-to-video tools. Executives referred to OpenAI’s Sora — first made publicly available not even a year ago on Dec. 9, 2024 — as a tool for pre-visualization. Agents casually described it as a creative reference point. Post houses quietly used it for filler shots and mood reels.
As with all of the large language models and video models bearing down on the entertainment industry, early adopters and optimists (myself included) took the view that this tech could enhance storytelling without completely supplanting storytellers. After all, in that long-ago time of a year ago, Sora was viewed as a clever brainstorming tool that could democratize content by making it easier for industry outsiders to visualize projects.
Related:
Everyone was experimenting, and few in Hollywood were admitting how quickly we were growing dependent on Sora (and tools like Sora). But on Sept. 30, when OpenAI released Sora 2 — the version that could drop real faces and voices into photorealistic, viral-ready videos — the industry’s quiet fascination snapped into open panic. Overnight, the film business saw what it had been enabling: a machine that could generate entire performances without paying, crediting or even asking the people it copied.
It wasn’t an innovation anymore. It was a mirror — showing Hollywood the endpoint of its own obsession with automation, franchise mining and cost-cutting. And that’s why, after years of rationalizing disruption as inevitable, yes, the industry is finally treating this as existential.
The Sora backlash isn’t just about IP or consent. It’s about survival. If image and performance are no longer human, then neither is Hollywood.
Today, I’ll lay out what you need to know about GPT-5’s far more dangerous OpenAI sibling — and how Sora 2 is pushing Hollywood to the brink.
What to know:
The leap that changed everything — Sora 2’s new video engine can now fake humans and performances so perfectly it’s poised to break the internet and the industry
Fake-out: How OpenAI’s “opt-out” policy left studios and stars scrambling to protect their own faces
The unexpected Hollywood powerhouse that finally said no — and forced Sam Altman to blink
How the debut of AI “actress” Tilly Norwood poured gasoline on an already-raging fire
The complicity: The uncomfortable truth — how Hollywood helped train the very machine now coming for it
A rallying cry: How this rollout is helping the entertainment industry find new solidarity and moral purpose
The solution: Yes, there is one — it’s going to take a lot of lawyers a lot of time, but here’s why it can’t wait