AI Risks, Human AI Agency, and Democratic Governance of AI

Artificial intelligence (AI) has emerged as a general-purpose technology with transformative implications across societal sectors. This working paper examines AI at four interconnected levels: societal risks, democratic implications, human learning and agency, and democratic governance. The central question concerns what AI means for human agency and judgement and for society's capacity for democratic governance. The working paper introduces the concept human AI agency, defined as individual’s capacity to engage with AI-systems critically, reflectively, and responsible while retaining human judgement. It further explores democratic AI governance. Citizens require human AI agency to participate in democratic decision-making regarding technological development, while democratic governance provides the conditions necessary for human AI agency to flourish. The working paper concludes that the future social effects of AI will depend on strengthening both human judgement and institutions capable of guiding the development of AI based on democratic principles. The overall argument draws on Giddens' theory of structuration.