Artificial Stupidity (Ma)
In The You On AI Field Guide
Ma's analysis grew from empirical work on AI adoption in organizations, where he observed a consistent pattern: deployments initially framed as augmentation drifted toward replacement as operators discovered cost savings from eliminating human judgment. The drift followed predictable incentive structures — quarterly reporting rewards visible cost reduction, while the costs of lost judgment manifest slowly and are hard to attribute.
The replacement pattern produces harm to organizations (decisions that worsen over time as contextual sensitivity is lost), harm to customers (whose problems receive algorithmic rather than situated response), and harm to employees (whose expertise becomes economically marginal). The enslavement pattern adds a specific form of worker alienation, in which human operators become extensions of algorithmic systems they cannot influence or override.
The concept explicitly extends Cipolla's framework into organizational theory, demonstrating that the quadrant structure applies to institutional actors as cleanly as to individual ones. The Research Society of Australia's parallel concept of Artificial Banditry covers the bandit quadrant; Ma's Artificial Stupidity covers the structurally more dangerous category of AI deployments that produce mutual loss.
Origin
Ma, a professor at Peking University, published 'Artificial Stupidity' in Long Range Planning in 2024. The paper synthesized organizational case studies with Cipolla's theoretical framework, producing one of the first rigorous academic extensions of the Cipolla laws to AI deployment.
Key Ideas
Two failure types. Replacement eliminates human judgment; enslavement dehumanizes human operators.
Institutional drift. AI systems deployed for augmentation migrate toward replacement under ordinary market incentives.
Triple harm. Organizations, customers, and employees are damaged simultaneously — the signature of Cipolla's stupid quadrant.
Cipolla confirmed. The empirical pattern validates the framework's applicability to technologies its author never encountered.