This Data & Society webinar marked the launch of Professor Michele Gilman’s report “Democratizing AI: Principles for Meaningful Public Participation.” During the conversation, she advocated for the need for a shared concept of what “public participation” means: without it, we may end up with models that are very tokenistic or ritualistic or even exploitative of the public.
In her policy brief, “public participation” consists of measures that offer opportunities for people most likely to be affected by the systems, to have influence in the systems design, and deployment, including decision-making power. This is not new to other regulatory sectors, such as land use, anti-poverty programs, and environmental law, all areas where the risk of harm is most serious for marginalised people.
According to Gilman, hard law requirements could help us push through the hurdles of public participation, by not leaving it to the benevolence of corporate or government entities. In this sense, a statute should define (1) the goals of public participation, (2) two-way deliberation between the public and decision-makers, and (3) affirmative actions to impacted communities, i.e. going out to communities as opposed to just asking for input online.
Harini Suresh, Assistant Professor of Computer Science at Brown University, shared the example of the “Data against femicide,” in which civil society activists took it upon themselves to collect and maintain the most accurate records of femicide in their contexts, given the lack of data, aiming to bring legitimacy and visibility to the problem. This initiative is an example of how having activists shaping the process and conceptualization helps build trust in long-term relationships, enabling solutions to be context- and community-specific.
Presenting a private sector perspective, Richard Wingfield, director of Technology and Human Rights at BSR, stated that they don’t normally use “public participation” as a concept, but the ideas behind it seem to overlap with the “stakeholder engagement” requirement under the UN guiding principles on businesses and human rights: “To enable business enterprises to assess their human rights impacts accurately, they should seek to understand the concerns of potentially affected stakeholders by consulting them directly in a manner that takes into account language and other potential barriers to effective engagement. In situations where such consultation is not possible, business enterprises should consider reasonable alternatives such as consulting credible, independent expert resources, including human rights defenders and others from civil society” (p. 20). Nevertheless, Wingfield highlighted that some places and communities lack the infrastructure, institutions, and social and political context to make meaningful participation happen.
The overall insight of the webinar was that this topic is not new, “people need not reinvent the wheel here,” which got me thinking about how current participatory practices can account for the rapid timeframes and new power dynamics posed by technology.