AI Games Are Taking Over Steam: But What About Quality?
There has been a massive increase in games using artificial intelligence on the Steam platform. The number of such games has exceeded 8,000, representing an 800% increase over the previous year. This means that one in 14 games on Steam uses this technology, and the share of newly released games featuring AI is expected to reach 20% by the end of 2025. This trend is profoundly changing the game development ecosystem, where AI now spans the entire industry chain. Approximately 60% of AI applications focus on generating virtual assets, such as character modeling or scene creation. Other common uses include sound synthesis, script generation, marketing material creation, and even code writing. Some developers are experimenting with real-time content generation, where AI dynamically creates game elements based on player commands or automatically detects prohibited content uploaded by users. This structure resembles a pyramid, ranging from basic asset generation through intelligent systems to fully AI-driven interactive experiences.
Experienced Developers' Perspective
However, this growth is generating differing opinions among developers. According to a survey conducted at the GDC 2024 conference, 52% of developers confirmed that their companies use AI tools, while 27% explicitly reject them. AI addresses certain pain points in development, such as rapid testing and iteration, thereby improving the overall quality of games in less time. For example, AI can simulate hundreds or thousands of players playing simultaneously, helping to uncover bugs, balance gameplay, and analyze real-world data for more natural animation and physics. This leads to greater realism and immersion, with games adapting to players in real time and generating new environments, quests, or customized stories. Visually unique AI-generated assets enable rapid experimentation, making games more engaging and personal.
Nevertheless, challenges affecting quality are emerging. Game development is undergoing a process of "de-technicalization," in which tasks such as modeling or programming can be completed using clever prompts alone, leading to a new form of competition known as "prompt competition." Many homogenized AI games are appearing on the market with similar styles and gameplay, differing only slightly in their prompts. This is causing an influx of soulless "quickly made" works and a shortage of talent for complex game genres. AI often lacks intent, so generated content can be incoherent, surreal, or unoriginal because it learns from existing data and copies older titles. Without human oversight, consistency declines, and many developers find that manually editing AI output takes more time than traditional methods. Steam's AI disclosure policy has gaps—the voluntary mechanism means that the actual number of games using AI is probably higher, while the unclear definition of the "extent of use" makes it difficult for consumers to assess.
As a result, AI improves quality in terms of speed, realism, and adaptability, but without human creativity, it leads to mediocre, derivative titles. The best approach is to combine AI with human creators, balancing technological innovation with artistic expression. Amid this industry shift, developers face the challenge of addressing ethical controversies and the boundaries of AI applications, while the market is flooded with games whose quality depends on how well AI is integrated into the creative process.



