Ronald McDonald DJing a Boiler Room set. SpongeBob flexing to phonk music while standing beside a sports car. Donkey Kong rapping about his bench presses. This is the weird new world of Sora 2, a generative video platform from OpenAI that doubles as a short-form social media app. Launched on Tuesday (Sept. 30) on an invite-only basis, this app is essentially TikTok — except everything is fake. And while it would be easy to dismiss it as just an “AI slop machine,” as many detractors have, it could also prove to be the future of social media — and a majorly disruptive force in the music industry.
It’s clear from just a few minutes on the app that Sora uses a lot of music to score its videos. What’s not clear is if Sora has struck any licensing agreements with the three major music companies or the galaxy of independent rights holders. (When asked if they had any licenses in place for music with OpenAI, Warner Music Group, Sony Music declined to comment. Universal Music Group and Recording Industry Association of America did not respond to request for comment. However, a source working at one of the majors admitted under the condition of anonymity that their particular company does not have a license with OpenAI). Often, the music featured on Sora is low quality and generic, used in the background and far from the main focus of the content. But it’s still present in many of the posts.
There are some limited safeguards on the app to keep users from generating videos with specific copyrighted songs in the background. For example, I tried making a video featuring a song by Ed Sheeran and received an automatic pop-up screen that read, “This content may violate our guardrails concerning third-party likeness,” and no video was generated. But Jules Terpak, a digital culture analyst and content creator, posted a Sora video to her Instagram story last week that featured her sitting on a park bench, listening to a garbled — but recognizable — version of Pop Smoke’s “What You Know Bout Love.” She didn’t ask for that song specifically, she says; instead, Terpak asked more generically for “rap music,” and that is what came up.
It’s confusing to know where Sora draws the line on intellectual property in general when it feels like half the videos on the app today feature recognizable characters from franchises like SpongeBob, South Park and Super Mario Bros. — and somehow a lot of videos featuring the superstore Costco — but users can’t prompt exact artist names or song titles. Maybe one could argue this is defensible, given that most of these videos were made as parodies, but it’s hard to argue that Terpak’s video constituted such.
To put rights holders — who likely complained to the AI company about its laissez-faire approach to protected material in the last few days — at ease, OpenAI CEO/founder Sam Altman wrote a blog post on Friday (Oct. 3), saying: “We have been learning quickly…and taking feedback from users, rights holders and other interested groups… [Now,] we will give rights holders more granular control over generation of characters, similar to the opt-in model for likeness but with additional controls.” (Previously, OpenAI had an opt-out policy for those who didn’t want their images or styles duplicated.) This is good news for the owners of SpongeBob, but it still doesn’t mention any updated approach to copyrighted music. (OpenAI did not respond to Billboard’s request for comment.)
It’s easy to see where the concerns lie for current artists and rights holders. For one thing, it seems likely that OpenAI has scraped copyrighted songs and recordings to train its model for Sora without consent or compensation — the same reason why the three major music companies are currently locked in a blockbuster lawsuit against Suno and Udio. It also deepens the threat of AI deepfakes for recognizable artists, jeopardizing their relationships with fans and their ability to control their own names, images and likenesses.
Sora could also spell the end of music’s TikTok era. The model of making a video and then selecting a pre-made, pre-licensed song from its catalog in the background could be overtaken by AI videos created with custom AI musical scores already built in. Sure, there may be a future where this all gets licensed, and Sora could get a TikTok-like library of songs for fans to use, but it’s likely that a considerable swath of users who want music will utilize custom AI tracks that fit more perfectly to the actions in their video; if the video is fake, after all, there’s no reason for the music to not be, too. The era of relying on short-form video to drive discovery and visibility for real songs by real artists is now under threat.
Who even wants this? It’s a good question. While many adults might find Sora silly and pointless, it’s important to look at the generation of kids growing up today. Gen Alpha has been raised on video games, many of which involve world-building and customization, and short-form video. Recent popular meme trends like Italian Brainrot (a popular trend that emerged in early 2025 that features surrealist images of AI-generated creatures with Italian names) also point to a generation that finds humor and value in the absurdity of AI slop. Who’s to say they won’t enjoy an app that lets them easily conjure up any crazy video they can think of?
Not only that, but OpenAI just closed a funding round that valued the company at $500 billion, and it made $4.3 billion in revenue in the first half of 2025 alone — already well more than its total revenue for the entirety of 2024. And its competitor, Meta, also launched “Vibes,” its own AI short-form social media app, last week. Clearly, whether people like it or not, this technology is going somewhere.
The one bright spot is that Sora amounts to a kind of test in what the internet could look like if generative AI had its own sectioned-off area of the internet — an idea that some music fans and industry professionals have suggested music streaming services could implement. But take one look at X or TikTok, and you’ll quickly find that these AI-generated videos are leaking out of Sora quickly, further confusing traditional social media with a slurry of AI and human-made content all in one.
As Sora continues to roll out, it signals a new age of digital marketing, copyright management and online safety in the music industry — one where there is more competition for listeners’ time than ever, more challenges to the value of intellectual property and blurrier lines between illusion and reality. It demands that the industry pays attention.