Altman says Sora could spark creativity. But is it just another addiction machine?
OpenAI insists Sora is a creative leap, but its timing tells another story. With billions in cash to burn and investors like Nvidia waiting for returns, the TikTok-style video feed looks less like art and more like OpenAI’s fastest path to ad dollars—and possibly its riskiest gamble yet.
Americans living next to vast artificial intelligence (AI) data centres now know what their higher utility bills are paying for: a new social-media time suck. OpenAI’s Sora is essentially the AI version of TikTok. Scan your face and record a few seconds of your voice and then use text prompts to generate videos of you—or a highly realistic, AI generated avatar of you—jumping out of an aeroplane with parakeets or dribbling a soccer ball on Mars.
If that sounds to you like a step toward dystopia, you’re not alone. The backlash to Sora this week was swift and brutal, calling out OpenAI’s hypocrisy in pledging to cure cancer but launching a trough for AI slop instead.
Sam Altman responded in classic slippery fashion on X: En route to noble goals, it was “nice to show people cool new tech," he tweeted. It’s certainly nice when you’re on course to burn through $115 billion in cash over the next few years, or when new investors like Nvidia would like to see a return on the $100 billion pledged towards your growth. Sora probably offers the clearest path yet to OpenAI hoovering up some advertising revenue.
While it’s easy to dismiss Sora as another form of AI brain rot, it could go as viral as ChatGPT. OpenAI has released it at a time when Facebook users are sharing overly glossy images of elderly couples and when AI videos of babies accidentally getting on aeroplanes and then piloting them are getting hundreds of millions of views on Alphabet’s YouTube.
Don’t ask me why people love this stuff, but it’s the reason Meta tried to land grab the AI video space last week with Vibes, a similar AI-video feed. While OpenAI’s Sora seems more social and meme-worthy by featuring users or their friends, Meta’s AI videos are longer and come from more creators. The goal is largely the same: emulate TikTok’s enormous success with an artificial twist.
Both approaches could gain traction. Meta has years of experience optimizing content for eyeballs, while OpenAI’s videos are technically superior, able to simulate real-world physics. Better character consistency also puts Sora ahead of Google’s Veo3—until now seen as the best AI video generator—which could produce that all-important social-media stickiness.
There’s something to Altman’s claim that Sora could spark creativity. It’s far quicker to post videos on Sora than it is to film and upload them to TikTok, where passive scrolling is pervasive. OpenAI says it has learnt from the costs of social media and designed Sora as a place to create and post rather than get addicted. To that end, it says that users will have greater control over Sora’s feed, to make scrolling a more positive experience.
The prevailing question on social media has been ‘Why has OpenAI released Sora at all?’ The answer is plastered with dollar signs: Generating millions of AI videos will be computationally expensive for OpenAI, but it could also build on the momentum set by TikTok and the trust created by ChatGPT to become a popular new platform for socializing and entertainment.
OpenAI now has product-development royalty in its executive ranks who will almost certainly turn this video feed into a natural space for ads. Doing so would be far easier than trying to insert contextual advertising into ChatGPT.
But the more obvious costs go beyond AI chips. There is a Sora video floating around which shows CCTV-style footage of Altman stealing, ironically, Nvidia chips from Target. That this was posted by one of OpenAI’s own engineers—who presumably had the boss’s consent to use his likeness—shows the level of disconnect that enthusiastic developers have when they throw transformative tech into the world.
Sora requires users to get consent from a friend to generate their likeness, but what happens when they create videos of people they know—or people of a different ethnicity—committing crimes, and start posting them on X and elsewhere?
That’s why public rebellion against Sora is so heartening. New epithets against AI chatbots like ‘clanker’ shows people have developed a healthy scepticism around how AI should be plugged into their lives, largely thanks to the shadow cast by social media. Young people are especially aware; 48% of teens say that social media harms people their age, up from 32% in 2022, according to a Pew study published earlier this year.
Altman’s claim of a ‘Cambrian explosion’ rings hollow because any tool built on the perverse incentives of social media is not truly designed with creativity in mind, but addiction.
Sora may spark a new wave of digital expression, but it’s just as likely to entrench the same attention economy that’s warped our online lives already. ©Bloomberg
The author is a Bloomberg Opinion columnist covering technology.
topics
