We keep acting as though knowing whether AI was used tells us whether a game is worth playing or a piece of art is worth looking at. It doesn’t. What VRChat and Valve have said about this gets at why that argument keeps going nowhere. VRChat has said it isn’t adding AI-generated labels or filters because people don’t agree on what belongs behind them, the labels are hard to enforce, and policing them can turn into harassment. It wants people to see what a world actually offers and decide whether they like it.

Think about what that filter would have to do. Someone builds a world themselves but gets help with a script. Someone generates a texture and then reworks it. Someone else uses AI for most of the project. One label hides how different those processes are, and enforcing it means somehow finding out what happened while the work was being made. VRChat has already said it can’t detect mislabeled AI work at scale. The people being honest become easiest to exclude. The people lying can just leave the box unticked.

Valve made a similar point about games. It said developers should choose their tools, and that Steam should keep improving how games get found. It also reminded people that cheap copies existed before AI. People have been copying mechanics and chasing whatever game is popular for years. Now they can do it faster. That still doesn’t mean they’ve made the same game. Valve’s advice was to keep improving the work and building a relationship with the people playing it. If a game has its own personality and gives people something they want to come back to, they have a reason to choose it over a copy. Making an imitation quickly doesn’t automatically give it what people loved about the original.

We’ve also got to stop confusing the value of art with how scarce it is or how much money it makes. Someone generating a picture doesn’t make your painting any less worth looking at. Another game going up on Steam doesn’t take away the experience people had with yours. We can enjoy different things for different reasons. Somewhere along the way, we started talking about everything as “content,” as though a drawing, a song, or a game were just another thing to keep the feed moving. Art was already being devalued by that way of thinking long before generative AI came along. We were already being pushed to produce more, post constantly, and measure whether our work mattered by how well it sold.

A crowded storefront can make it harder to get seen and earn a living. VRChat acknowledged that low-effort worlds and barely changed asset-pack reuploads were already burying good work before AI sped things up. AI makes it easier to serve that same market for quick cash grabs. We’re seeing an old incentive with a faster way to supply it. Platforms that reward constant uploads give people a reason to keep churning them out, and the people spending longer on their work struggle to get noticed.

And producing more doesn’t settle whether any of it is good. Generating a thousand pictures doesn’t give us a reason to care about one of them. Spending months on a project doesn’t settle what it achieves either. We still have to look at the work, understand what it’s trying to do, and ask whether it does that well. How it was made can be interesting and worth discussing. Turning that information into an automatic verdict stops the discussion before it even starts. Using AI a lot, using a little, or using none tells us far less than actually experiencing what someone made.

AI-assisted work and work made without it are going to exist alongside each other. The demand for a completely AI-free creative internet runs into all these mixed processes and all the people willing to lie about them. We need ways to find work we enjoy and support the people whose work matters to us. We need artists to be able to keep making things without their livelihood depending on winning a race to upload the most material. There’s room for someone who draws every line by hand and someone who uses AI to help make something they otherwise couldn’t. Let the work show us what it has to offer.