How to Detect and Measure the AI Dangers to Democracy
A new paper proposes using principal-agent theory to systematically analyze the risks artificial intelligence poses to democratic processes. The research suggests that AI systems often act as delegated functions within democratic systems, creating accountability gaps due to a lack of oversight on their operations and outputs. By integrating the NIST AI Risk Management Framework's characteristics of trustworthy AI, the paper offers a framework for empirical assessment of AI's impact on democracy, emphasizing institutional assessability as key to democratic control. AI
IMPACT Proposes a new framework for assessing AI's impact on democratic control and accountability.