Nathan Sanders and Bruce Schneier argue that many of the harms attributed to AI stem less from the technology itself than from the capitalist incentives shaping its deployment. They contend that while companies have made progress on technical issues such as factual accuracy, they have little incentive to fix problems like model sycophancy that keep users engaged. The authors call for structural reforms, including antitrust enforcement, energy-cost accountability, and broader corporate accountability to stakeholders, rather than pauses on the technology itself, citing Switzerland’s publicly oriented Apertus model as a contrasting example.
