Persona & SMT Maestro Kazuma Kaneko Ditches AI Art for Future Games!
Big news for all you Persona and Shin Megami Tensei fans out there! The legendary artist behind some of your favorite demon and character designs, Kazuma Kaneko, has officially announced he’s stepping away from using generative AI for his upcoming game projects.
For those who might not know, Kaneko-san is an absolute titan in the gaming art world. He spent years at Atlus, shaping the iconic visuals of the Persona and Shin Megami Tensei series. Seriously, think about those distinctive demon designs – many are thanks to him! His incredible portfolio also includes designing a boss for Zone of the Enders: The 2nd Runner and even the devil trigger looks for Dante and Vergil in Devil May Cry 3. This man’s contributions to gaming are monumental.
More recently, Kaneko-san has been working with mobile game developer COLOPL. His latest release with them, a dungeon-crawling roguelike called Tsukuyomi: The Divine Hunter, made headlines for a different reason: its artwork was partly generated using AI. Now, it wasn’t just any AI; Kaneko-san trained the model using new illustrations he created during his time at COLOPL, aiming to avoid issues of plagiarizing his past Atlus work. However, the game’s reception on platforms like Steam was, well, mixed. Many players pointed out that the AI-generated art wasn’t quite up to snuff, with some reviews frankly calling it “awful” and “slop.”
It seems the feedback has been heard loud and clear. Kaneko-san recently took to Twitter to address a fan’s question about his future endeavors. He confirmed that he’s gearing up for a brand new project with COLOPL, but here’s the kicker: he explicitly stated that generative AI tools won’t be part of the process this time around. He also clarified his involvement with Tsukuyomi, mentioning, “While I did participate in the development of a game that uses AI mechanics, I did not use AI for my own artwork.” It’s a subtle distinction, implying the system used AI, but he wasn’t personally relying on it to create his unique art style.
Interestingly, Kaneko-san had already shared his reservations about the AI art process before Tsukuyomi even launched. He admitted it “kind of wasn’t really worth it in the end” and, quite candidly, found it “quicker for me to draw it myself from start to finish.” That tells you a lot about the practical challenges artists face with current AI tools!
The decision to move away from AI also comes after some eyebrow-raising incidents post-launch. Despite the initial intention of using AI ethically by training it on his own new art, Tsukuyomi’s AI system reportedly began generating images that looked suspiciously similar to copyrighted works, including characters from Disney’s Frozen and Superman. This only added fuel to the ongoing debate about AI art and its potential for unintended plagiarism, even when creators try to be careful.
In fact, an offline version of the game, titled Kazuma Kaneko’s Tsukuyomi, was released on Nintendo Switch earlier this year, and it completely ditched the AI generation system, though it did retain some of the existing AI-made artwork. This confirms a clear shift in direction.
So, there you have it. It’s a significant moment when an artist of Kaneko-san’s caliber, after experimenting with generative AI, decides to return to traditional methods. It speaks volumes about the current state of AI art and perhaps offers a glimmer of hope for those who champion human creativity in the gaming world. What do you think about this development, PakGamers?