Is AI The Future of Gaming, Or Just Overhyped? An Expert Weighs In

Artificial intelligence in gaming is a hot topic these days, with everyone from developers to the average gamer wondering what it means for the future of our beloved pastime. Is it the dawn of a new era, or just a lot of buzzwords that won’t deliver? To get to the bottom of it, we turned to a seasoned pro: Luke Dicken, an industry expert with a long history in AI research for companies like Take-Two and Zynga, who was exploring AI long before the current generative AI boom.

Dicken believes firmly in AI’s power as a tool to enhance games, but he’s quick to differentiate between the useful applications and the current wave of hype. For him, AI is essentially “anything that allows a machine to make a decision.” The real magic happens, he says, when AI can do things either “faster than a human could make it, or when you need more of something than a human could make.” Think about games like No Man’s Sky, with its impossibly vast, unique universes – that scale would be impossible for human developers alone. Or remember Left 4 Dead’s brilliant AI Director, which dynamically adjusted everything from zombie spawns to weather, creating that constant, nail-biting tension. These are examples of AI truly serving the game design.

However, Dicken is far less convinced by the flashy demos we’ve seen of large language model (LLM)-backed dynamic NPCs, like those shown by Nvidia. He points out that while they offer more choices, more choices don’t automatically make them more interesting. The crucial question, from a design perspective, should always be: “What does this allow us to do that’s beneficial to the player experience and that we couldn’t already do before?” If it’s not genuinely enhancing the player’s journey, then what’s the point?

When it comes to using generative AI to automate developer roles or churn out game assets, Dicken sees potential, but with a massive caveat. He acknowledges there are good applications, especially for handling non-creative, statistically average tasks that need to be done efficiently. But the idea of generative AI as a “magic button” that automates entire development pipelines? That’s pure fantasy, he stresses. The true value of AI, he argues, isn’t in doing old things a bit faster or cheaper, but in tackling scale issues that humans simply cannot manage.

At its core, Dicken believes that “technology has to be in service of art to be effective.” He feels this fundamental principle has been largely ignored amidst the current generative AI frenzy. He finds it “exceptionally weird” how a few companies have managed to stir up so much debate and even economic concern over what he considers “marginally interesting algorithms.” He compares generative AI to scaffolding – it can support construction, but it can’t build a house on its own, and you definitely wouldn’t want to live in one it tried to build.

Adding to his concerns is the economic reality. Dicken highlights that Big Tech companies are currently burning massive amounts of money to subsidize these generative AI systems. The “math just isn’t mathing,” he says. These tools, which are already expensive even with subsidies, could see astronomical price hikes – potentially 10 to 20 times, or even 1,000 times in extreme cases – once the subsidies dry up. This, he warns, is a “five-alarm fire” for businesses, many of whom have been lured by low initial prices without truly quantifying the return on investment (ROI) of their AI implementations.

Ultimately, Dicken doesn’t see generative AI as the silver bullet for the AAA gaming industry’s notoriously high costs. Instead of looking for shortcuts through AI, he suggests publishers should focus on the widespread problem of wasted labor – the countless projects that go through multiple resets, pivots, and leadership changes, often for years. He questions how much of the “ballooning strained cost” actually goes into building the final game, rather than being eaten up by internal inefficiencies. For Dicken, the real solution isn’t just about faster tools, but smarter leadership and more efficient development practices.