Sora 2 helped me make a number of videos no human crew could have produced. One was a ten-shot sequence from a sci-fi spaghetti western, featuring a duel between an aging weathered sheriff (generated from my selfie) and an armed anthropomorphic aardvark. Sora clones your voice, and the realism is shocking.
Another Sora 2 video I created was a dreamlike return to the split-level suburb of my childhood. The third showed me winning the World Series. None of it was real, of course, but my emotional response to the videos sure was, and it wasn’t just me. A friend saw the baseball clip on social media and commented. “I know it’s fake, but I felt happy for you.”
Hollywood keeps saying AI is just another tool, like computer animation or digital editing. That is a comforting thought. It suggests the business can absorb the change the way it always has, now with an AI-improved business model. But OpenAI’s Sora 2 is not just a cheaper way to make special effects or do previsualization. It is the first version of something that changes what audiences expect a story to be. Until now, movies have been a shared experience. You watched actors embody a character. You projected yourself into the scene. AI breaks that contract. Now the audience can star in the story.
The Aardvark Western was the most traditional of the ten and fifteen-second videos I prompted with Sora 2. I wrote a shot list, including dialogue, pacing, and camera framing. It took several attempts, but Sora eventually delivered the sequence I imagined. I felt like a director. Then I could easily change the aardvark to a porcupine, a giraffe, an alligator, etc. One prompt, many versions. Endless variation without additional cost. “Content is liquid,” as Matthieu Lorrain of Google DeepMind frequently says.
The second video, “I’m Your Son,” came from a much looser prompt: “Take me back to the split-level suburb where I grew up.” Sora filled in the rest. In the first shot, a glowing blue glowing portal (straight from “Back to the Future,” and “The Terminator”) brought me to the past in just a step away. It’s cold. I’m wearing a winter jacket. You can see my breath. It looks like my memory, but not exactly. It was a machine’s reconstruction of suburban Americana, assembled from millions of images. The emotional effect was real, though the details were invented. It felt like someone else had dreamed of what my past might be and added a dramatic twist. Nostalgia is something Cinematic AI does well, because it was trained on every movie ever made.
The third clip was the most revealing. Sora showed me a younger-ish (42 instead of sixty-something), stronger, better me, triumphantly blowing a 99 mph pitch by the last batter to win the World Series. Everyone goes nuts. A childhood fantasy fulfilled. It is obviously fake, yet the reaction was joy. The point was not authenticity; it was wish-fulfillment. A synthetic moment was able to deliver a feeling that normal entertainment cannot provide. A Hollywood movie lets you watch someone else win. AI lets you win.
The change is that now viewers can appear in their own stories. The fantasy once reserved for actors, athletes, and celebrities can now be generated on demand. Once people get used to that, it will be hard to return to the old deal. Movies will not disappear, but they will no longer be the only dream in the machine.
We are about to live in a media environment with three lanes: stories made by humans for a mass audience, stories made by humans for short-form platforms, and stories made by AI for a single person. They will coexist for years, maybe decades, the way film, television, and YouTube still coexist. But the hierarchy changes. A story no longer needs a cast, or a greenlight. It only needs a viewer.
In a world full of synthetic faces, real humans become more valuable. Celebrity has become a luxury category. The more AI content floods the zone, the more people will pay for the assurance that a person with a name, a history, and an ego created something. In the same way vinyl became premium in the age of iTunes, real actors may become premium in the age of generative media. As synthetic content spreads, real people become more valuable. Celebrity shifts from ubiquity to scarcity. Authenticity turns into an even bigger selling point. Today, it is the celebrity system that defines Hollywood, not the former studio and network monopolies.
Human creativity remains, joined by machine-made work. Stories for mass audiences, short-form videos, and individualized AI stories will exist side by side with live-action ones. Some people will prefer the easy, unearned nostalgia of AI stories. We already make these choices in food, art, and news. These media choices will vie, as they do today, for our narrowing attention spans and increasingly individualized palettes.
Header image credit: Zac Wolff on Unsplash
Charlie Fink is the author of the AR-enabled books “Metaverse,” (2017) and “Convergence” (2019). In the early 90s, Fink was EVP & COO of VR pioneer Virtual World Entertainment. He teaches at Chapman University in Orange, CA.