Take-Two President Says Google’s Genie Won’t Replace Dev Workflows

Take-Two Interactive president Karl Slatoff dismissed comparisons between Google’s Project Genie and traditional game engines, stating that generative artificial intelligence models cannot replicate the full scope of game development. He made the remarks during the company’s latest earnings call, shortly after Google expanded access to its experimental world-generation platform. 

While the publisher sees value in AI tools, it maintains that creative production remains human-led.

Google’s Project Genie, powered by Genie 3 alongside Nano Banana Pro and Gemini, is described as a general-purpose world model capable of generating interactive environments from text or image prompts. The prototype web app allows users to sketch, explore, and remix “living” virtual worlds in real time, simulating physics and interactions as players move.

Early tests spread quickly online, including unofficial recreations resembling major franchises. Some users also generated experimental scenarios, such as a fan-made GTA-style setting. Investor reaction followed. Take-Two shares fell roughly 10%, Roblox dropped more than 12%, and Unity Software declined over 20% after the reveal.

Project Genie AI vs Traditional Game Engines

Slatoff stated the technology is being overstated compared to established development tools, stressing that “Genie is early in its iteration” and “not even in the same ballpark” as traditional engines. He added that it is “not a game engine,” contrasting the prototype with production platforms that manage core systems such as gravity, lighting, audio, physics, and gameplay mechanics.

Yet, Slatoff stressed that environment generation is only one layer of production. “There are so many more elements to game development that go beyond world creation,” he stated, citing narrative design, emotional engagement, mission structure, and creative direction.

Project Genie currently functions closer to a procedurally generated interactive video than a full game experience, with roughly 60-second sessions at about 720p and 24 fps and limited gameplay. Generated worlds lack objectives, sound, persistence, and stable continuity, according to The Verge.

Despite the caution, Take-Two continues to integrate generative AI internally. CEO Strauss Zelnick said the company is “actively embracing” the technology to drive efficiency and innovation, with the long-term aim of expanding creative possibilities rather than replacing developers.

Project Genie remains an experimental research prototype rolling out to Google AI Ultra subscribers in the U.S., with broader access planned. For publishers like Take-Two, the technology signals workflow evolution, not the end of traditional game studios.