Containing AI: from responsible to critical AI governance

Abstract This paper argues that most responsible AI (RAI) governance frameworks do not meaningfully address AI’s environmental and social costs. Rather, they merely create the impression of managing these problems. With the example of “sustainable AI”, we show that, in its most extreme form, current RAI governance not only results in weak regulation but in a regulatory vacuum . Against this, we propose the concept of critical AI governance . With this, we reframe the governing question from “how do we make AI responsible?” to “do we need AI at all?” This approach rests on three overarching principles: (1) critically questioning dominant ideologies and economic imperatives; (2) prioritizing social and political solutions over technological fixes; (3) investing in public digital infrastructure and independent research. Together, these contributions point toward a different style of AI governance altogether.