Strauss Zelnick made some comments about using AI in video games, and if you didn’t realize it, it should have raised your eyebrows.
As reported by Rockstar Intel, this came from a paywalled interview with CNBC’s SquawkBox. Rockstar Intel was asked about using AI in Grand Theft Auto 6, and this was what Zelnick said:
“We’ve been using digital tools for since the very beginning. I would say the interactive entertainment business in an innovator within what people call AI now. I don’t doubt we’ll continue to innovate. I think it is worth noting though the genius is human. The tools might be digital but the creative genius is human.”
Zelnick’s comments come across as positive and particularly upbeat on the video game industry. But we have to point out that Zelnick is conflating at least two very different things, that may be called AI, but at least most people understand are different.
Artificial intelligence has been in development since the 1950s, and also went through boom and bust periods as a field of research. A 1980s boom led to Japanese investment and application in robotics. In 1997, Deep Blue beat world class chess champion Garry Kasparov, both a symbolic and technical achievement.
As Zelnick correctly points out, the video game industry has always been using AI, and in fields like enemy and NPC behaviors, procedural generation, and computer generated graphics, it has innovated and pushed the envelope in using AI.
But Zelnick didn’t talk about the kind of AI that the public is worried right now. AI model technology is a particular kind of AI that uses a large amount of data to generate different kinds of outputs. The public best knows generative AI itself, as we saw an influx of images and videos generated with AI online.
This technology may use some older technologies like deep learning and neural networks, but it’s very new. In fact, the first identifiable AI model tech, 15.ai, came out only one year before it went mainstream with OpenAI’s DALL-E in 2021.
Here and now, most gamers are concerned about this new technology being used to make AI images in video games. We have reported on Bungie, Ubisoft, Electronic Arts, and Activision talking about or being caught using generative AI or AI model tech.
On one level, some gamers genuinely dislike the aesthetics of AI generated art. Beyond the simple errors of hands sometimes appearing with six fingers instead of five, not everyone likes how these images and videos look.
On another level, the public has become rightly concerned that AI models have a huge carbon footprint. Subsequently, their proliferation has multiplied the carbon cost the video game industry is spending on the planet.
On top of that, many gamers are worried that the use of this technology is harming the jobs of artists and creatives in video game companies. If these artists are not being fired, are they getting paid and compensated less because of them? And on another level, are they being hampered creatively as their companies decide to use more AI?
And connected to all these points is that video game companies are perceived and observed to not be disclosing their use of AI properly. We cited Activision not being transparent above, but we must credit EA for going the opposite route.
They explained in detail they created their own AI model from scratch, called Opta, for EA Sports FC, and its application did not come at the expense of their developers, especially their artists and creators.
And that’s our concerns with Zelnick’s statement. As Grand Theft Auto 6 has yet to release, many fans have real concerns if Rockstar used AI model tech, if this led to game developers losing their jobs, etc. There’s no question a multibillion dollar production like this will have a huge carbon footprint, but depending on how AI was used, that footprint could have ballooned exponentially.
We certainly hope, if not Zelnick himself, people at Rockstar can reassure us on this issue. It will be ugly if we have to wait for the game to release to find out the worst.
