Although many game studios today have begun to use AI tools in the development process, the vast majority of these tools are far from practical value. This perspective comes from Jon Gibson, head of transformation at Keywords Studios, a leading provider of technology and creative solutions to the gaming industry.
Keywords Studios is headquartered in Ireland and was founded in 1998. It has provided technical and development support for many well-known works such as "Dune: Awakening", "Marvel Contest", "Call of the Abyss", "Alan Killer Remastered", "Dead Island 2", "Marvel's Guardians of the Galaxy" and many other well-known works.
Gibson said in an exclusive interview that the Keywords team tested the applicability of more than 500 game development AI tools, and only about 6 of them had real practical value.
“There are a lot of flashy products out there,” Gibson said. “Keywords has done a lot of internal R&D work. We tested about 500 different AI tools in multiple R&D projects, and ultimately felt that only about 6 could really help development work in a reasonable way.”
He explained that as far as the current status of the industry is concerned, the development logic of most AI tools is centered around "what functions the tool can achieve" rather than "what specific problems in game development can be truly solved." This ultimately leads to the fact that these tools may achieve some "cool-looking" effects, but have no practical application value.
"The current AI technology is in a chaotic stage of wild growth, and we need to push it into a stage where it can be truly implemented." He further explained, "How do we apply AI in actual production environments? How do we make AI a boost to the development team rather than a potential threat? At the same time, how do we use AI within a normative framework to ensure that it is in a controllable state, ensuring intellectual property security, legal compliance, and ethical and moral legitimacy?"
"There is a huge gap between those cool AI demonstrations that can generate stunning effects by inputting a prompt word, and AI applications in actual production environments where humans control the direction and output stable and high-quality results. It is not easy to cross."
He added: "Many people only focus on what looks cool and focus on the tool or model itself, rather than what they really want to achieve. Some companies will use or even develop a tool without clear application scenarios, and then forcefully stuff it into the existing production pipeline. The correct logic should be the other way around: first clarify 'what are our pain points and what problems do we want to solve', and then develop targeted tools."
Gibson also mentioned that in the latest industry status report released by the Game Developers Conference (GDC), 90% of the developers interviewed said that they have used AI in development, but at the same time, 52% did not recognize the application of AI.
"90% of people are using AI, but 52% think it is a bad thing." He concluded, "There is obviously a serious gap in the middle. I believe that the lack of control and standardization system for the application of AI is one of the core reasons for this problem. Many companies have not really explained to employees why they use AI, what the value of AI is, and what the relevant strategic planning is. Without clarifying these core issues, rashly enabling AI models will naturally make developers worried."