Democratic Governance Of AI Systems And Datasets

Progress in AI’s capabilities has highlighted a significant concern: the uneven allocation ofAI’s benefits among and within societies and economies, worsening both domestic and globalinequalities. In addition to limiting market competition and innovation, this concentration ofpower also translates into social challenges: global inequalities, risks of bias and disinformation,replacement of jobs, and lack of democratic control. The Hiroshima AI Process G7 Digital &Tech Ministers statement of 1 December 2023 states that one of the key goals of the processis “to maximize and share the benefits of this technology for the common good worldwidewith partners beyond G7”. The policy interventions outlined in this brief serve this purpose,by supporting democratization of AI technologies. Ensuring democratic governance of AI canlead to greater oversight and control of AI technologies and systems, levelling the playing fieldand making AI technologies more accessible, and ensuring that AI deployment respects humanrights and democratic values. And at global scale, ensuring democratic governance of AI meansaddressing concerns for global digital divide and inclusivity, building local capacity and engagingstakeholders in the global majority countries. The brief outlines a set of policy interventionsfocused on the provision of resources necessary to democratize AI development and use,through digital public infrastructures and digital public goods provided by G7 member states.This includes computing power, training datasets and open-source AI solutions. Additionalrecommendations concern managing AI-induced labour market changes and ensuring aframework for international and multilateral cooperation.