When Openai released its new AI-generated video app Sora last week, it launched with an opt-out policy for copyright holders—media companies would need to explicitly indicate that they did not want their AI-generated characters running rampant on the app. But after day Nazi SpongeBob, Criminal Pikachu, and Sora Philosophizing Rick and Morty, OpenEE CEO Sam Altman announced The company will reverse course and “let the rightsholders decide how to proceed.”
In response to a question about why Openai changed its policy, Altman said it came from talking with stakeholders and suggested he did not expect the outcry.
“I think the theory of how it was supposed to make people feel, and then actually seeing the thing, people had different reactions,” Altman said. “It felt more different from the images than people expected.”
The TikTok-like Sora app offers an endless scroll and the ability to create 10-second videos—with audio—of virtually anything, including your own AI-generated self (dubbed a “Cameo”) and anyone who has consented to use your likeness. Although it tries to restrict depictions of people who are not on the app, its text prompts have proven more than capable of generating copyrighted characters.
Altman said many rightsholders are excited, but they want “a lot more control,” adding that Sora “became very popular very quickly … we thought we could slow the ramp down; that didn’t happen.”
“We obviously really care about rightsholders and people,” he said. “We want to build in these additional controls, but I think you’ll see that many major pieces of content will be available with restrictions on what they can and can’t do.”
Among early adopters of the system, Altman said he was surprised that people would have “in-between” feelings about allowing people to create AI-generated videos using their likeness on Sora. He said he expects people will either want to try making their Cameos public or not, but it’s not like there’s going to be much nuance to this, and that’s why the company recently introduced more restrictions. Many people have made up their minds about whether or not they would like their Cameo to be public, Altman said, but “they don’t want their Cameo to say offensive things or say things that they find deeply problematic.”
Bill Peebles, head of OpenAI, Posted on x Sunday that the team had “heard from a lot of people who want to make their Cameos available to everyone, but maintain control over how they are used,” adding that users can now specify how their Cameos are used via text instructions to Sora, such as “Don’t put me in videos that include political commentary” or “Don’t let me say this word.”
Rightsholders want ‘too much control’ over Sora
Peebles also said the team is working on ways to make the Sora watermark on downloaded videos “clearer and more visible.” Many have raised concerns about the misinformation crisis that could naturally come from hyperelastic AI-generated videos—especially when watermarks characterize them as AI-generated and can be easily removed, according to video tutorials proliferating online.
“I also know that people are already looking for ways to remove it,” Altman said about the watermark during a Q&A Monday.
During Altman’s keynote at DaveDay, he said the company was immediately releasing a preview of Sora 2 into OpenEye’s API, allowing developers to use the same model that powers Sora 2 and create ultra-realistic AI-generated videos for their own purposes, without any kind of watershed. During a Q&A with reporters, when asked how the company will implement security measures into the API for Sora 2, Altman did not specifically answer the question.
Altman said he was surprised by the amount of demand for creating videos just for group chats — that is, to share with just one other person or a handful of people, but not more widely than that. Although it’s popular, he said, “it’s not a great fit for how current apps work.”
They positioned launch speed bumps as learning opportunities. “Not much longer, we’re only going to have one good video model, and there’s going to be a ton of video with any of our security measures, and that’s OK, the way the world works,” Altman said. “We can use this window to society to really understand, ‘Hey, let’s change the playing field, now you can be prepared in some cases.’ ,
Altman said he thinks people don’t pay attention to OpenAI’s technology when people at the company talk about it, only when they release it. “We’ve got … this kind of technological and social co-evolution,” Altman said. “I believe that works, and I don’t really know anything else that works. There are obviously going to be challenges for a society that struggles with this quality, and what will get better with the video generation.
“Clearly there are going to be challenges for society with this quality.”
This one is controversial, especially for AI CEOs. For as long as AI has been around, it has been used in ways that impact minorities and vulnerable populations – in the wrong way arrests To AI generated RogueOpenai has some guardrails for Sora, but if history – and the past week – is any guide, people will find ways to get around them. Watermark removers are already proliferating online, with some people using “magic eraser”-type tools and others coding their own ways to explicitly remove watermarks. For now, text prompts don’t allow generating a specific face without permission, but people have allegedly already got around that rule to generate close-enough approximations of someone to cause fear or intimidation, and to create suggestive videos, including videos of women holding objects such as dildos.
Asked if Openai’s plans translate to a “move fast and break things” approach, Altman said, “Absolutely not,” adding to that user criticism of Sora right now, saying the company is “way too restrictive” and “censorshipful.” He said the company was starting the rollout conservatively and would “find ways to allow more over time.”
Altman said turning a profit on Sora “wasn’t in my top 10 concerns, but … obviously someday we have to be very profitable, and we’re confident and patient that we’ll get there.” He said the company is currently in a phase of “investing aggressively”.
Early Difficulties Greg Brockman, president of OpenEye, said he was struck by the adoption curve for Sora – and that it was even steeper than chat. It has been at the top of the list for free apps in Apple’s App Store. “I think it points a little bit to the future — this is the thing we keep coming back to: We need more computation,” he said. “To some extent, this is the number-one lesson of the (Sora) launch so far.”
This was essentially a pitch for Stargate, OpenAI’s joint venture with SoftBank and Oracle for AI infrastructure in the US, starting with a $100 billion investment and adding up to $500 billion over four years. President Donald Trump has championed the venture, and OpenEAI has announced a handful of new data center sites in Texas, Ohio and New Mexico. energy-hunger-projects has been controversial, and they are often able to run with the staff couple hundred people Despite promises of massive job creation after initial construction,
But Openai is full speed ahead. On Monday, it struck a deal with chipmaker AMD that could allow OpenAI to take 10 percent stakeHours later, in a Q&A with reporters, Altman was asked how interested the company was in making its own chip. “We are interested in the full stack of AI infrastructure,” he responded. At another point, he told reporters they should “expect to hear a lot more” from OpenEI on the infrastructure stack.
During the session, OpenAI executives repeatedly stressed the lack of computation and how much it could block OpenAI and its rivals from being able to offer services at scale.
“Asking, ‘How much calculation do you want?’ Is it like asking, ‘How much of the workforce do you want?’ ‘Brockman said. “The answer is you can always get more out.” And right now, more ability to deepthroat your friends is the latest selling point.
