Slay the Spire 2 makes placeholder art a statement against generative AI

Slay the Spire 2 shipped its early access with placeholder art that looks deliberately rough — crude Microsoft Paint-style portraits mostly drawn by co-founder Casey Yano. The move reads at first as a rejection of the industry's growing habit of using generative AI for placeholder assets. In an interview with GameSpot, Yano confirmed that sits behind it, but the real reasoning is more layered, and it is about protecting how a game is judged, understood, and eventually made.

Much of the "shit art," as the studio itself calls it, lives in the game's Timeline section, which grows as players unlock epochs and carries the game's overarching narrative forward. Yano's first argument is about expectations. A deckbuilder roguelike lives and dies on balance, and early access depends on players understanding the game is unfinished. "I wanted to set the precedent that some of the stuff is incomplete, and I wanted to make it obvious," he said. "It has to look like shit. It's important that it looks like shit." If the art looked close to final, he argued, players would judge the game as a finished 1.0.

Legibility matters just as much. The team prioritized which assets needed to be finished so the game stays readable in play — a relic that is just a red box, Yano noted, "conveys nothing," and players need to know which relic will trigger. But the crude drawings are also a working framework, not just a placeholder to be thrown away. When placeholder art passes from Yano to the art director Marlowe Dobbe or another illustrator, the roughness signals intent: "If you draw something a little too crude, they'll be like, 'This must be what Casey wanted. This must be intentional.'" So the drawings deliberately carry composition, color, and detail that the final art gets built from.

The strongest argument Yano makes is about authorship. Speaking "as someone who draws," he objects to generative AI because it removes authorial intent: "When you draw anything, you draw a thousand lines. Each line is not some serious decision, but there's a decision process for nearly every line." He notes the distinct look of AI-generated art — "very dense, very colorful, and very distinct" — and asks why a studio would need AI "to mimic our own style" when the game is already in their style. He also points to what generative output erases: "You don't see the path that somebody took to get better and develop their own style."

The placeholder pictures also carry the personality of the studio. Yano described "the playfulness of development," and told a story about a younger developer named Jose whose placeholder art is full of mechas and pop-culture references — "some Among Us, some SpongeBob" — that got pulled once early access made the game more public. Looking ahead, Mega Crit wants players in on it too: once Slay the Spire 2's content settles, the team plans a community contest inviting fans to draw "janky" art for some monsters, added as optional in-game art. Yano's one reservation is that fans might use generative AI for the contest, which would defeat the whole point.