Generative AI is a contentious hot topic in the gaming industry, and understandably so—its impact has been disruptive, either driving up hardware prices or causing job losses among professionals. Although there are some seemingly practical everyday applications, the hype surrounding the technology and its poor public relations image are hard to ignore. Executives constantly boast about AI creating Matrix-like worlds, and people are long since tired of hearing these empty promises.
So, when speaking with Tim Willits, Chief Content Officer at Saber Interactive and co-owner of id Software from 1995 to 2019, this topic naturally came up. At Gamescom 2026, he told Joshua Vallance of PC Gamer, “As everyone knows, we’re experimenting with AI technology in our future games. We didn’t bring any related titles to this year’s show, but we previously announced Ride-Hailing Simulator, and I’d like to clarify a few things, as there are quite a few misunderstandings circulating.”
You might be wondering why we’re discussing a ride-hailing simulator game. Here’s some background: earlier this month, Saber Interactive found itself embroiled in controversy over this title. The issue was that the game hadn’t clearly disclosed its use of AI technology on Steam. This came to light after a former writer exposed the situation, claiming they were removed from the development team because the company intended to replace human labor with AI.
Saber Interactive’s CEO, Matthew Karch, initially publicly disparaged the writer before later issuing an apology. Willits clarified that the main campaign and all side quests in Ride-Hailing Simulator were written by humans, with AI components used only for an optional game mode called “Free Ride”: “We use AI to generate passenger characters, allowing us to produce more new stories.”
“So, only AI can make this possible—it’s an experiment. The Saber team wants to explore how to use this technology, so we’re starting small to gain experience. Like others in the industry, we’re still figuring things out. But none of our major flagship titles feature generative AI. Hellraiser doesn’t, Turok doesn’t, and Stuntman doesn’t either.”
I think this claim that “only AI can achieve this” deserves some scrutiny. Admittedly, generating an infinite number of dialogues for an optional mode is indeed something only generative AI can do. But if the goal is simply to provide enough dialogue for a free-roam mode to keep it from getting boring, real writers and voice actors could easily handle that. Sure, content eventually runs out, but I doubt anyone expects Ride-Hailing Simulator to be a game they’ll play for a lifetime.
Despite this, Willits emphasized, “We want to run experiments, and I believe there will eventually be a company—I don’t know which one—that creates a game using AI so masterfully that everyone will say, ‘Yes, this is how AI should be used.’” He also joked, “Someone has to create the ‘Half-Life 2 of AI,’ a title so stunning it directly spawns killer applications.”
Personally, I remain as skeptical as ever. To be fair, though, I agree with Willits on one point: if generative AI truly becomes core gameplay, it will likely spawn entirely new genres rather than being used to create endless RPGs with infinite NPCs. That approach ultimately results in hollow, tedious, and meaningless experiences.
My skepticism stems from watching vast amounts of money burn up and data centers spring up, only for the resulting technology to seem useful for little more than menial tasks. Once AI becomes central to production, the voice acting ends up flat, the writing is a mess, and the art lacks both style and memorability—no matter how advanced the underlying technology becomes.
If Willits’ so-called “Half-Life 2 of AI” actually materializes, I’ll eat an AI-generated hat served with a side of humility. But too many people have already tried and failed to reach this goal, and I feel it’s becoming increasingly out of reach. Nevertheless, Willits stated that if generative AI had been available back then, id Software would have certainly used it—because the studio has always been at the forefront of technological innovation.
“Of course! Think about how id Software used to operate. When we released Quake III Arena in 1999, the game required a dedicated graphics card, and we got so much flak for it. People said we were ruining the entire industry, that no one would spend $100 on a graphics card, calling it the stupidest decision in history... I’m certain that if this technology had existed back then, John Carmack would have definitely created something like Skynet.”
To be fair, however, Willits seems to share my view: games need to prove through their own quality that generative AI can transform the industry, rather than relying on empty promises from executives. “The game itself will speak volumes. Take our Ride-Hailing Simulator experiment, for instance—I encourage everyone to try it out after launch. They might think, ‘Wow, this is cool, I like it.’ If they don’t, that’s fine too. You’ll know whether the game is good once you’ve played it.”