Debates about art are also useful to the people who own the platforms

These conversations can be like [art-washing] – they draw our attention the way that murals draw our eye to the street-level view of new developments while the decisions about what to build and how go on somewhere else. Usually those decisions have been made before the mural even goes up. Whatever we think of the building or the mural, debating whether we like the mural is not the same as debating whether the building should be there.

Whether we think machines can themselves make art, the more we debate whether they will ruin or end art, the more time we spend assuming that they have the power to do so. This [frames] the discourse in a way that ultimately serves the purposes of anyone who would benefit from people thinking image generators make art – or are simply more powerful or intelligent than they really are. [Timnit gebru] has done a lot of work to show the same thing about AI-alarmism. If you get a lot of people frightened of the idea that an intelligent machine could take over the world, you get a lot of people thinking our machines are powerful enough to do that. When we debate the artfulness of “AI” art, we participate in the broader projects of branding machine learning as intelligent, developing plain-language interfaces as products, and staking claims to broader conversations about artificial intelligence.