Jon Lai of Riot Games and A16z Opens Up About AI In AAA Games
As we reported back in the summer, AI is making inroads into gaming. This was seen with the release of the very first completely AI-created video game. According to the creators of This Girl Does Not Exist, the entire game from music, character, art, and more was generated using a variety of AI-powered tools. As many know, machine learning and AI use to create video game elements isn’t new. From the earliest inception of gaming, AI has been utilized as a tool. But now, major studios are paying more attention to technology as they look toward the future. One clear example of this is Activision/Blizzard’s copyright for an AI program that plays customized music for players.
Because of this and other advancements, it seems that many within the industry are paying attention and are ready to discuss the ramifications. Jon Lai, former Product Manager of Riot Games, wrote a lengthy Twitter thread on AI and AAA games. His focus is the cost of producing high-quality AAA games and how AI could allow smaller studios to produce the same quality titles for a fraction of the cost & time. His prime example is CD Projekt Red’s Cyberpunk 2077. In this example, Jon broke down the overall costs in capital and time to get the game from concept to ship. The short of it was, “500 developers over 9 years and ~$175M to ship.” These are numbers that match Hollywood blockbuster movies.
He went on to discuss “the triangle theory.” There is scope, cost, and time. Teams can pick just two. The reality according to Jon Lai would be as follows. “Building a big, immersive MMO with a small team? Expect it to take 10 years or speed it up by hiring more but it’ll cost $$$.” This concept is common in other industries as well. Each of these factors determines costs so teams must choose which two are most important.
But this is where artificial intelligence comes in. With new AI-powered tools, the triangle is broken as it’s currently known. That’s because emerging tools are promising to drastically cut down the price of scope, speed, and time. Jon Lai returns to Cyberpunk 2077 and provides the following example. “Imagine Cyberpunk where much of the world / text was generated, enabling devs to shift from asset production to higher-order tasks like story-telling & innovation…all while shipping in 3 years vs 9.” And, as Jon Lai points out, the technology for text generation is already here thanks to ChatGPT.
But that’s not all. He goes on to detail multiple facets of game creation, from creating NPCs (Non-Playable Characters) using AI tools such as PrismaAI, to genAI which creates 3d assets that many games such as Diablo use. All in all, these tools are proving to be effective at reducing costs in terms of labor, and power. For small indie studios with limited manpower, this could become a game changer, allowing smaller teams to produce AAA quality titles without needing a multi-billion dollar company. If you’re interested in this subject, I strongly suggest that you read the entire thread. It’s a major eye-opener and Jon does an amazing job at crunching the numbers.
The thing is though, it’s not just gaming where we’re seeing AI being utilized in this manner. As ODSC reported over the last few months, restaurants are looking to AI to enhance labor. From AI answering systems, to real-time inventory controls to manage rushes, fast-food and sit-down establishments are investing in AI-powered tools. It’s becoming clear that a driver for the adoption of AI tools will continue to be, the reduction of costs.
Originally posted on OpenDataScience.com
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday. You can also get data science training on-demand wherever you are with our Ai+ Training platform. Subscribe to our fast-growing Medium Publication too, the ODSC Journal, and inquire about becoming a writer.
