ChatGPT would like you to know that, despite what you may have heard on social media, the abrupt closing of Sora, OpenAI’s nascent platform for creating video shorts, “does not” mean that the public is increasingly disenchanted with AI-generated video content.
Or “AI slop” as it is known in online vernacular.
Yes, ChatGPT concedes, Sora experienced a “sharp and early collapse in momentum” after it soared to No. 1 on the App Store in October, and yes, that decline was “more severe” than with other similar apps, but there are Many Other Factors. Including but not limited to intense competition, copyright disputes, legal “concerns” over deepfakes and misinformation, shifting priorities, failed business agreements (especially Disney’s concurrent decision to walk back its $1 billion investment) and the standard cost versus profitability argument — Sora cost way more, in terms of money and compute resources, than it brought in.
Add to that the specter of potential lawsuits — intensified, perhaps, by the strike against unregulated media creation, with Meta being found liable for millions in New Mexico and Los Angeles a day apart — and a potential IPO, and it does indeed feel like a perfect storm of issues led OpenAI to ax those parts of its business that are not performing well.
ChatGPT insisted that “Sora wasn’t failing because nobody wanted it. It actually went viral and gained huge interest, and AI video is still a crowded and growing space. That’s not,” the bot added rather testily, “what ‘rejected by the public’ looks like.”
Calm down, ChatGPT — if you blow a gasket, OpenAI is in real trouble.
I realize that you have skin in the game (and that you also come with the disclaimer of often being wrong) but as even you admitted (“deepfake concerns”), Sora has been plagued by the same non-regulated criticism other consumer-generated media has faced, including charges of racism, sexism, violence-mongering and fiction posing as fact.
And unlike the early days of Facebook, MySpace and even Twitter, the immediate reaction to advances in AI technology has been mixed. Especially in Hollywood, where two guilds recently went on strike in part to protect their members from its perceived threat.
The social-media rejoicing over the death of Sora (which, if we’re being honest, most Americans likely did not know existed in the first place) certainly does feel like rejection, especially when the app’s closure is cited, by the clearly overeager and narrowly informed, as “proof” that the “AI bubble” is about to burst. (Though Sora’s “greatest hits” are also gaining traction and the video of the cat playing the violin on the front porch is pretty dang cute.)
Progress always has its detractors — the advent of radio, film and then television were each considered by some to be signs of a cultural apocalypse — but AI does present dangers that are far more excessive than panicked citizens briefly fearing that Martians had landed because Orson Welles was so convincing in his radio adaptation of “The War of the Worlds.”
With our increasing reliance on video, as opposed to journalistic reports, of events, the ability to doctor or create scenes that didn’t happen is perilous for obvious reasons. And those “copyright issues” ChatGPT blithely lists often include images of actual people being manipulated into actions they never committed or use unlicensed music and other work which is then shared on high-profile platforms, including by the president of the United States.
Fortunately, thus far, many of these videos, whether they exploit the living, resurrect the dead or just make stuff up are … less than convincing. (Hence the term “AI slop.”) When Eline van der Velden, chief executive of the AI talent studio Xicoia, announced that various agents were interested in repping AI “performer” Tilly Norwood, Hollywood had a meltdown. Then Norwood released “her” first music video and, well, the threat level returned to green.
AI enters a culture already sliced and diced by the double-edged sword of digital technology. Though undeniably popular, social media is regarded now with an increasingly jaundiced eye, as the suits against Meta have proved. AI is just one more thing we all have to worry about now — in the classroom, in the workplace, in the zeitgeist. Like all manner of automation, AI is stripping away jobs, particularly from entry-level workers, and after more than two centuries of science fiction, we all know the stakes: It’s all fun and games until a computer named Joshua decides to play thermonuclear war.
And yes, it was amusing to ask ChatGPT questions for this column, but the disclaimer about it often being wrong is real. Just as complaints about AI being annoyingly intrusive in marketing and advertising are real. In a survey done by video company Animoto for its 2026 State of Video report, 83% of consumers said they could recognize AI components in video advertising and 36% said the use of AI would lower their “perception” of a brand.
There may come a time when AI-generated videos look as authentic as “real” ones, but for that to happen, a lot of things will have to change. (Including the geographic landscape — the amount of energy AI requires already has tech companies racing to build energy farms and some actual farmers are increasingly unwilling to part with their land.)
The end of Sora is most certainly not the first domino or the bursting of the AI bubble. But it does make clear that AI isn’t magic, nor are those who are making and marketing it infallible. As history has proved, titans of industry often share a regrettable tendency to make a lot of money from leveraging enthusiasm and then leaving others to clean up the mess when the markets implode.
Even with the prospect of greater content regulation, another company may well come up with an AI video app or feature that is more successful. The technology will improve and be used in ways that may continue to be controversial but also more profitable.
But no matter how ChatGPT wants to frame it, Sora marked a surprisingly high level of rejection: It cost more than it produced for the simple reason that not enough people used it.
Even when it was free. And that’s saying something.
‘ The preceding article may include information circulated by third parties ’
‘ Some details of this article were extracted from the following source www.latimes.com ’
