AI in game development remains a sharp topic that bothers many users. While some developers believe that it is a good tool to optimize workflow, others categorically discard the idea of using artificial intelligence in their projects.
Furthermore, projects become advertised as fully handmade, without the use of AI. But even these factors don’t help with the audience seeking traces of artifacts in games. This happened with Crimson Desert in March 2026.
After the release of Crimson Desert, the game peaked at more than 245,000 concurrent players, making it a large and successful project from the get-go. However, with more players joining, users started to notice various artifacts on the screen.
This way, players noticed strange textures, deformed models, and, overall, generated images in in-game paintings. It sparked a large wave of discussion, since the project has never disclosed its use of AI, especially on Steam, where it remains a mandatory condition for game publishing.
The main points of criticism remained not only the generative AI usage, but also the lack of transparency during the game’s promotion.
The response to the criticism was quick – developers from Pearl Abyss reacted to the controversy and acknowledged the use of generative AI. They stated that this was not a calculated, misleading move but a simple development mistake that made its way into the final build.
Initially, AI was used as a way to create placeholders and concepts for earlier stages and world-building designs. However, during production, assets were not adapted or changed, but left at a final build as a mistake.
To fix the issue, developers promised a more thorough analysis of all the assets in the game and changes of all the AI-generated models. As for today, the audit is still in progress, since no new announcements were made.
Such a case sparked a new wave of discussions about generative AI. Going back to controversies with Larian and Sandfall Interactive, we can see that not all the developers are strictly against the AI usage. For example, Swen Vincke, the CEO of Larian, stated that using AI for concept art is not the same as replacing artists, but rather an opportunity to give them tools.
On the other hand, Tyler Glaiel, the co-developer of Mewgenics, was not so sure and, regarding the issues with Crimson Desert, stated that concept art is “the worst place to use AI.” David Szymanski, developer of the Iron Lung, also stated that even though the AI future is inevitable, he is not ready to welcome it at any capacity.
Meanwhile, other game developers, such as Pavel Konstantinov, CEO of RetroStyle Games, stated the following:
“AI usage is a controversial topic for many. For example, in my studio’s pipelines, we implement AI solutions to boost the production speed. However, the main point is not to replace artists but to give them a new tool for convenient work. While it provides a significant boost, AI still requires professional developers for further processing of the results.”
From the retrospective, it becomes obvious that AI usage will be further explored in games. Even looking at the commentary of industry experts, we can see that the opinions divide. Using generative AI even as a tool at pre-production stages remains a dividing point for many.
However, with more loud and popular examples of proper AI utilization, the industry might change its perspective on the matter. The loudest cases mainly concern mistakes and low-quality assets left in the game. With that, we can predict that more attention to final builds and enhanced QA will help developers use AI advantages without compromising results.
