When Everyone Can Create AI Video, Does Security Become the Real Battleground?

When Everyone Can Create AI Video, Does Security Become the Real Battleground?

How many software releases have there been with regard to AI video production software? An (instantly out of date) list would include the latest Veo, Sora, Higgsfield, Nano, Seed dance etc. I’d estimate my production company Synima use around 400 different bits of software on a regular basis. Maybe it’s more – certainly new players enter the market all the time. And companies gobble each other up of course.

Now all of these are improving at a ‘hockey stick’ rate, and yes that’s astonishing. It means that anyone can create a wide variety of AI generated video and animation. But before you think I’m going to outline some argument for and against AI video production, there’s another angle to this.

Thought experiment. This time next year there will be Veo 5 and Seed dance 12 or whatever. It will be incredible and do amazing things and even more people will be able to use it. Character consistency will be nailed, everything will be more realistic etc. So: more ability to produce content for more people. Probably some agentic AI going on in the workflow. And that equals more ‘slop’ or more amazing work depending on what happens. But perhaps that’s not where the competition will be.

If everyone can produce the content then won’t the next frontier will be protecting the workflow and ultimately what is produced?

So a brand that wants a video produced will easily be able to get that video made. No problem. But what are the security issues around it? How ‘walled’ is the workflow? How exposed are they making themselves to reputational damage?  How protected is their IP while they are making it? How tight is the NDA with the freelancer they’ve employed? How likely is it that the company they’ve used to make it won’t subsequently be bought by another company? Who owns the copyright of the prompts and final content, Higgsfield or you?

So then I think.

Get insurance? It’s slippery.  Try getting a number on what your IP is actually worth?  A simple brand pallet, or an animated brand ambassador like ‘Felix the Cat’ (Purina)? I’ve had ‘what is theft?’ when it comes to trade secrets and ideas. ‘What is ‘theft’, bit semantic? And that’s not helped by the US Supreme Court that said a week ago that ‘entirely AI created content cannot have US copyright protection’. In any case who want’s to lock the door after the (copyrighted) horse has bolted? 

In house Production.

Well lots of companies are setting up production internally and can easily set up a ‘walled garden’ of their own. But that won’t work. For starters, how can any production company work without the cloud these days? But more generally, most of my experience working with in house production companies is that they’re full of inefficiencies the vast majority of which are not of their own making. For example take procurement. They are effectively hobbled by procurement who won’t purchase what they want when they need it. Procurement will say ‘not this budget cycle’, ‘vendor not on approved list’, ‘due diligence get three more quotes’ whatever … by the time the in house team have the software the world has moved on. So what’s the chances of their  security protocol being air tight when it comes to agentic AI worming it’s way in? What happens if a bad actor decides they want to gain access to the IP of that in house production company? There have been some stunning corporate breaches recently: JLR, Co Op, and in our specific industry Larson, Toei, and again, Sony.

So while the conversation rages on about creativity (Netflix acquisition of Ben Affleck’s production company anyone?). And when the President of The United States randomly posts clips of ‘Iron Man’ to support his ‘Epic Fury’,  it’s worth revisiting what will make a great production company in 2028 and beyond. It may actually be something as creatively ‘boring’ as security.

As usual, I’d love to know your thoughts.