Acceso usuarios
Adele Barzelay, Jeremy Ng, Mila Romanoff
With the global growth of AI applications and automated decision-making systems across public and private sectors, countries globally are grappling with how to govern AI responsibly and encourage AI innovation to promote the Sustainable Development Goals (“SDGs”), while still protecting key rights and interests. While national, regional, and global governance models are starting to emerge, it is too early to assess trends that could work as “good practice models.” The sheer volume, scope, and pace of AI regulatory efforts across the globe, and especially in the Global North, may pose a challenge for Low- and Middle-income countries (“LMICs”) seeking to address the risks of AI in their local contexts. While trying to develop appropriate governance frameworks for rapidly developing AI technologies, many LMICs still face gaps in access to digital infrastructure, broadband connectivity, electricity, availability of quality data sets, digital literacy, and overall capacity.
Such digital divides risk being compounded by a global AI divide, as a handful of private sector actors in the developed world continue to entrench their control over key data and AI infrastructures.
In seeking to avoid mimicry and enable sustainable development and technological innovation, what should LMICs consider in developing their AI governance frameworks? Through our analytical and operational experience, we have learned that Low- and Middle-income countries design and enforcement of effective governance approaches and institutions in LMICs requires considering the specific context, values, social, economic, political landscape as well as capacity and resource constraints. LMICs may have the opportunity to engage in more “greenfield” approaches to governance, given the relative lack of legacy policy, regulatory, and institutional frameworks. This Paper will seek to provide an overview of some of these tradeoffs, as well as a structured approach for LMICs to consider putting in place the building blocks of an AI governance regime that can support sustainable development by fostering “trust.”
