Last week, OpenAI, the company that brought you ChatGPT, introduced Sora, an artificial intelligence model that can turn text prompts into realistic videos. As I watched them in amazement, I thought of the many creatives whose jobs will be in jeopardy when this technology goes into widespread use — special effects people, visual artists, and so many more.
Sora even creates lifelike fake people. I see why SAG-AFTRA fought so hard to have wording in its new contract with movie and TV producers to ensure that images of its members couldn’t be rendered in AI-created scenes without paying the actors for their use.
I remember being amazed when I first saw Pixar’s “Monsters, Inc,” in which Sully looked as if he had real fur which moved around as the character did. It was a major breakthrough in computer-generated graphics, which probably took months for Pixar’s digital artists to create. But when you watch the sample set from Sora (which took considerably less time to render), one of them shows two woolly mammoths charging towards the camera, and their fur appears just as realistic. The same applies to the scene of a Samoyed and a Golden Retriever “romping through a futuristic city at night.” Only this time, the AI engine created those images, not humans.
That said, the software still needs a prompt before it can create anything. But whatever a writer can conceive of, it seems Sora can create. As the video below from the Wall Street Journal indicates, however, Sora doesn’t do everything perfectly. Yet. Remember, this is an early iteration. Like all the technology that preceded it — from the telephone to television to cars to light bulbs to the internet — Sora (and similar products sure to come) will only improve with time.
It won’t be long before all sorts of AI-generated videos of real people, places, and incidents are being created. That will open the door for even worse misinformation, blurring the lines between true and false more than anything we’ve seen before, endangering further our common knowledge base of what’s actually going on in our world.
There are no safeguards in place to prevent bad actors from creating absolute nonsense that looks even more convincing than today’s deep fakes. Imagine the impact it could have in this presidential election year, where we already have a right-wing media ecosystem that promotes lies and fictional memes around the clock.
Unfortunately we can’t count on Congress to regulate this technology, because its members won’t understand it. Despite background work and briefings from their aides, the senior citizens on The Hill have proven their tech ignorance again and again during hearings, as if they were grandparents who still can’t unmute themselves on Zoom.
I’d love to see what Sora could do with that prompt.