Tag
The article proposes a framework for quantifying system-level harms from AI adoption in complex sociotechnical systems, using a case study on adversarial manipulation in financial infrastructure to demonstrate increased risks.