One of the big problems with the tech industry's current obsession over large language models (commonly referred to under the umbrella term "AI"), is separating out the genuine potential applications and implications of this technology, from the marketing hype, misunderstanding, and outright lies proffered by industry leaders and blue-tick grifters pivoting from NFTs. A good example of how the two are easily conflated was recently provided by Activision CEO Bobby Kotick, who shared his thoughts on AI's potential applications in gaming.
As reported by Kotaku, Kotick was asked his opinion of AI tech in a company-wide meeting last week. "I don't know how much people realize that a lot of modern-day AI including ChatGPT started with the idea of beating a game, whether it was Warcraft or Dota or Go or Chess. But what is now these large language learning model AI technologies, all started from this idea of beating a game."
Presumably, Kotick is referring to specialised supercomputers like IBM's Deep Blue, as solving games and game-like puzzles were an early target for developing artificial intelligence. It's a perfectly reasonable comment, but it isn't especially relevant to what lies ahead for modern LLMs.
Kotick then expresses his belief that LLMs will be as influential as the "first MacIntosh" in terms of "how meaningful the impact of AI would be on society both positive and negative." He then adds, "For what we do, I think it will have a profound positive impact on the things we'll be able to do in game development for a long time. It will enable us to do things that we haven't been able to do for a long time."
Fair enough. But what things, exactly? Kotick uses Guitar Hero as an example. "I've always had this vision for what a new Guitar Hero product could be but without having AI and then the processors embedded either in phones, in computers, or game consoles that allow you to actually have the speed of processing to enable that AI, we've never been in a place where AI is going to have practical reality and applicability for games until now. And I think when you look out over the next five or seven years, the impact in game-making is going to be extraordinary."
I'm sorry, what? Kotick says he has a vision but never expresses what that vision is. Is it a Guitar Hero with AI-generated songs? A Guitar Hero that can autogenerate challenges based on existing songs? He doesn't specify. Instead, he just says that whatever the vision is, it wasn't possible until now. But possibly not now, but in five or seven years.
The exchange has the air of Kotick not being prepared for the question, and using a mix of passingly relevant historical knowledge and vague assertions to come up with an answer. But it's indicative of how the gold rush to profit from AI tech is muddying the waters regarding what the actual capabilities and ramifications of the tech are. Everyone wants to be in on it. Everyone wants to be seen to be in on it. Everyone's making big claims about how it's going to reshape society. Ultimately, nobody really knows what the future holds right now, and nobody in charge seems to care about pausing to consider it.
Well, almost nobody. For a very different perspective, check out the more measured AI take of Blackbird Interactive's CEO, who states "There's no AI-driven software that I know of we would put in a shipped game", and goes on to provide a very thoughtful analysis about the abilities and limitations of LLMs like ChatGPT.
