Pode a IA substituir governos eleitos por governação direta da IA dentro de 20 anos ?

Os sistemas de IA já são usados para analisar impactos de políticas e simular resultados económicos com crescente precisão. Alguns argumentam que uma governação conduzida por máquinas poderia ser mais eficiente e menos corrupta do que os sistemas liderados por humanos. Os críticos alertam que isso poderia levar à perda de responsabilização democrática e de agência pública. A questão de saber se os algoritmos podem realmente representar valores humanos coletivos permanece sem resposta. Será que o fim da democracia representativa já está a ser arquitetado?

Background

AI is already deployed to model policy outcomes and simulate economic impacts with growing precision; proponents argue machine-driven governance might curb corruption and deliver efficiency gains, while skeptics emphasize risks of democratic erosion and the unresolved challenge of encoding collective human values into code. Existing systems remain narrow administrative tools rather than sovereign decision-makers, and leading assessments caution against granting AI autonomous governance powers. Per the Stanford HAI 2024 AI Index Report, there is no evidence that AI could, within twenty years, design, deploy, and sustain governance capable of safely and legitimately substituting for elected governments across diverse societies; current AI lacks causal understanding, normative reasoning, and moral agency, and public opinion, legal frameworks, and constitutional constraints worldwide remain firmly opposed to unelected algorithmic rule. Current deployments focus on narrow administrative support rather than policymaking, underscoring the gap between technical capability and the institutional legitimacy required for algorithmic sovereignty. SOURCE: Stanford HAI — https://hai.stanford.edu/ai-index/2024-ai-index-report