RAND Corporation researchers Benjamin Boudreaux and Alvin Moon developed a mathematical model based on social choice theory to track how artificial intelligence could erode human agency over time. This gradual loss of control threatens collective decisionmaking through disenfranchisement, AI enfranchisement, and invisible agenda control. As automated systems increasingly shape the options presented to human decisionmakers, society risks crossing a critical threshold where independent human authority is quietly superseded.
To identify this tipping point, the model tracks key trends, including how often humans override AI recommendations, the scope of automated authority, and the time available for human deliberation. This insidious transition could become irreversible due to growing technological dependence, the high cost of removing embedded systems, and the loss of human operational expertise. To counter this trajectory, the researchers propose implementing "agency audits" to systematically evaluate how new models impact human decisionmaking authority before deployment.
No comments:
Post a Comment