AI specialists in the media sector give their reactions to the closure of OpenAI’ s Sora AI video generator

This morning, OpenAI abruptly pulled the shutters on its high-profile AI video generator, Sora. It closed down the Sora app and website without giving any reason, and shortly after Disney ended its deal with OpenAI which would have seen it invest $1 billion in the company and allowed its characters to be used in AI generated videos.
We asked several AI specialists in the media sector to give their reaction to Sora’s closure. You can read their comments and opinions below.
Benjamin Field, CEO and co-founder, Deep Fusion Films

There’s always been something intriguing about why Sora has never been publicly available in the UK, and you’ve needed a VPN to be able to use it here. I spoke to Chad [Nelson, creative specialist at OpenAI] about this and challenged him about whether it was about copyright. He told me specifically that it wasn’t, but it sounded like he was holding something back.
18 months later and it still hadn’t been rolled out in the UK, which made me suspicious that it might have been an IP issue after all. Especially as OpenAI has always been quite guarded about where its training data came from.
The material I saw from Sora was really strong but I really only saw the fringes of what we were allowed to see. It was technically very competent; the challenge is it was at odds with the industry’s calls for safeguarding and protections.
The latest release really seemed to be deeply problematic, with no safeguarding to protect people’s identity. You could make anything – even Sam Altman was in a video which was made without his involvement at all.
For any of these generative AI models to make money they need widespread adoption, and Sora seemed so far away from what the industry is calling for. There seemed to be an arrogance to it at a time when the industry is looking for data provenance, to safeguard human creativity, and to be paid for creative material that’s used for training.
The UK Government’s recent reversal on its outlook for training data, where ‘opt in’ isn’t the default, is important. I’m sure this didn’t change OpenAI’s mind, but it was probably waiting to see what the government was going to do, and its move has very slightly shifted the power dynamic back to human creativity.
How far OpenAI is behind others that have consenting material in their models, and how much it would have needed to invest in that material, could have been a big factor in its decision to close.
This is in contrast to Google, which, through YouTube, has a vast amount of material it’s allowed to use for training and that’s updated every day with tens of thousands more videos, all having implied consent to be used for training. That is the level of perpetual data you need, and those are the models that will survive.
Now it’s stopped Sora, I think OpenAI will pivot to real-world building, and we’ll see its video creation focus come back as a games engine or something close to that, where it could offer something that will add value to the industry.
FOLLOW AI MEDIA NEWS on LinkedIn for regular free stories about AI’s use in media and the creative industries

Matt Campion, co-founder, Spirit Studios
We’ve used Sora as part of our creative toolkit, so it’s definitely a shame to see it go because it opened up some exciting possibilities. But it’s also not the only platform doing great work, and the whole space is moving at a ridiculous speed.
From my perspective, OpenAI has been dealing with two big pressures. One is the growing online pushback from people who don’t love the idea of paying for tools from a company they see, rightly or wrongly, as connected to defence tech.
The other is that Google was always going to have a huge advantage simply because YouTube gives them this enormous, living training ground for Gemini.
So, Sora disappearing feels less like a dramatic collapse and more like a sign of how competitive and complicated this landscape has become.

Guy Gadney, CEO and founder, Charisma.ai
Sora had a great launch with much promise but had started to lag behind other video models such as Google and Seedance.
We believe that these models will have transformative impact effects on the TV industry, which is why we pioneered Charismatic so that UK creators can have more control of these developments rather than responding to overseas events.

Benjamin Woollams, founder, TrueRights
OpenAI has pulled the plug on Sora, and with it, the first real attempt at an AI-native social platform. It’s a significant moment for the industry.
I applaud the ambition - but the execution was rushed. They tried to drive relevance through equity deals with high-profile talent, which generated headlines but didn’t solve the core problem: most users still don’t understand why they’d use an AI video platform, or what it unlocks for them creatively.
Sora was built around novelty and gimmick, not genuine creative utility – and that’s a fundamental misread of where adoption actually is.
Social platforms require content, and content is fuelled by creators. Creators require compensation, or at the very least, recognition and attribution. Sora had neither. Without that engine, there’s no flywheel.
The bigger question now is what this does to OpenAI’s positioning in the content space more broadly. They made a bet, it didn’t land, and the market will remember that.
For Disney, this could end up being one of the smartest things that’s happened to them in the AI space – even if it wasn’t by design. The deal was never fully finalised, but what they did get was an invaluable window into market reaction, consumer sentiment, and the gaps in how a licensing agreement of this scale should actually be structured – and there definitely were some.
They come out of this better informed, better positioned, and with every option still on the table. If anything, this feels like a get out of jail free card.
The fundamental issue was always the deal structure itself. Exclusive licensing arrangements between studios and single AI platforms – even for a defined period – run counter to the direction the market needs to move. But beyond exclusivity, there was no technical infrastructure underneath this agreement. No guardrails on how IP was being used, no attribution framework, no enforcement layer. That’s not just a legal exposure, it’s a strategic one.
Without that infrastructure, Disney would have had no real-time visibility into how their characters were being used, by whom, in what context, or at what scale. That’s a loss of brand protection, yes, but more importantly, it’s a loss of insight.
No comments yet