GenAI is coming for your video…

The camera may never lie, but there's no cameras involved in this video. Seeing is no longer believing…

Adam Tinworth
Adam Tinworth

Well, last week OpenAI continue its assault on the creative industries by starting to preview its generative video model: Sora

At first glance the results are pretty impressive:

0:00
/0:10

That looks like the sort of thing that would have taken weeks of expensive CGI work to do normally.

0:00
/0:08

And using AI to create a shot like that is a lot cheaper than travelling to the location and putting a drone up.

It's an example of how technology makes ever cheaper:

  • Arial shots used to need a helicopter: very expensive
  • They they needed drones: buying a drone was cheaper than renting a helicopter once
  • Now AI can just fabricate the shot for you, for a few pennies…

Of course, these are carefully selected examples. Sora isn't open to the public yet, so we can't see what the success rate is on shots like this. Some of the examples CEO Sam Altman has been sharing on X are superficially impressive, but you can quickly see the tell-tale AI signs, like a suddenly appearing extra arm on the monkey…

I can see this sort of stuff being useful, especially for the sorts of social videos that are almost entirely built out of stock footage nowadays. But equally, this is just going to open up the floodgates of ever more compelling misinformation.

Oh, how quaint “seeing is believing” and “the camera never lies” seem these days…

AIvideo

Adam Tinworth Twitter

Adam is a lecturer, trainer and writer. He's been a blogger for over 20 years, and a journalist for more than 30. He lectures on audience strategy and engagement at City, University of London.

Comments