AI and the Threat of Stable Totalitarianism
Author: Keir Reid
Summary: This paper delves into the risks posed by AI in undermining democracies and emphasizing autocracies. It explores the concept of "lock-in totalitarianism" where certain values or practices become entrenched in a system, making it impossible to reverse. The paper highlights how AI can be misused by governments to surveil, suppress dissent, manipulate behavior, and control populations, ultimately risking the erosion of open societies. By pointing out recent developments in AI technologies such as automated censorship, metadata analysis, facial recognition, generative AI, and persuasive algorithms, the paper urges for collaborative efforts to mitigate the risks posed by authoritarian misuse of AI. It emphasizes the urgent need for a global response to the threat of stable totalitarianism and stresses the importance of factoring in these risks into AI policy decisions to safeguard democracy and human rights.