What’s OpenAI’s new Sora video generator going to do to games?

• Sora is the latest pitch OpenAI has towards their ‘transformation’ of the digital world
• All fears aside, what’s actually going to happen?
• And more importantly, why care?

If you’ve been watching the news recently it would have been hard to miss that OpenAI – the company behind ChatGPT – have debuted their newest tool, Sora, a ‘text-to-video’ generator that they showed off in a Twitter thread that drew reactions both impressed and horrified.

But, we don’t need to rehash the issues that AI presents again. Check out our other article for more information; but in brief, the legality of data these models draw from to generate their output needs tightening and replacing human artists through pulpy, pith action at speed is destructive. There are big promises, but even bigger fears of how it’ll denigrate the artistic process.

But outside of the fears about replacing components of development, art and whatnot, what else could these tools do?

The death of Kickstarter

For many developers just getting their start, creating their own games takes far more time and money than balancing another job would allow. So a lot of people turn to Kickstarter, where players can invest early with the promise of a full game down the line. Usually, some sort of proof-of-concept is needed if you want to elicit the most investment. But what happens when someone can just mock that up with AI, like in this Minecraft-esque generation shown by OpenAI below?

Okay, Kickstarter and other early-access frauds aren’t rare today anyway. The Day Before is a particularly infamous example. But quite often nearly all of these had to generate their – fake – gameplay by actually building it in an engine, even on a very limited scale. Players have become accustomed to looking out for more evidence to see that a game has promise, and CGI trailers are practically a red flag in and of themselves.

But if Sora’s promise of being able to make complex, realistic-looking footage – even of games – is true then we’re in for a very bad trip indeed. If not convincing then we’ll most likely see more than a few people try to change their arm and convince some kind-hearted players to fork over cash for their dream game with a seemingly provable concept video to go with it.

What else besides?

Aside from that well, we’ve always seen fraudulent adverts in mobile gaming. It’s the entire reason that they introduced legislation that mandated you had to have the actual gameplay shown featured in your game. So, surely, realistically we won’t see much difference on mobile, especially since – like we already said – you’ll have a hard time convincing hardcore gacha geeks who desire cool new characters for their unique art and abilities to pay for something that was AI-generated.

Why care?

But there’s one more important point to be made. Why care? The fact is that, like any new tech over the past few years this has all the hallmarks of big ideas and little to no practical implementation so far. Procedural generation this ain’t, and as anyone could point out the kind of things being made by AI so far would require serious thinking by a human third party to get working in a game engine. It’s all well and good showing some limited snippets, but none of said snippets really show anything but vaguely impressive yet oddly uncanny copy-pastes of what we’ve seen before.

Are there cool, ethical uses for AI? Maybe. But I don’t think that it’s all going to be good with the way it’s currently being approached. Yet I don’t think it’s apocalyptic either. I think that, like any changing technology, people will always value the human touch, and with so many of the best games on mobile having their own unique style or creative aspect then you won’t see Sora or Midjourney replacing them anytime soon.