The Berkman Klein Center, in collaboration with the City of Helsinki’s Education Division and the AI-transparency company Saidot, hosted a virtual AI Policy Research Clinic to study the ethical governance of artificial intelligence-enhanced technologies deployed to support learning, student wellbeing, and retention in Helsinki’s vocational schools. Part of a Global Network of Internet & Society Centers(NoC) project on the Ethics of Digitalization, the Research Clinic convened a dozen early-career scholars to examine a real-world use case of AI in the public sector. The larger initiative is conducted under the patronage of German Federal President Frank-Walter Steinmeier and is supported by Stiftung Mercator.
During the first portion of the month-long program, the Research Clinic explored questions related to the creation of an inclusive, participatory, and sustainable strategy for stakeholder engagement throughout the design, development, deployment, and assessment stages of the new technology. During the second half, the student cohort examined the viability and appropriateness of different human oversight mechanisms. In both instances, the primary goal was to create resources to help the City of Helsinki, as well as similarly situated municipalities, to navigate thorny AI governance challenges related to participatory design and human oversight.
Divided into two working groups, the first group produced four distinct outputs: a human oversight model to enhance cooperation across the City of Helsinki’s technical, operational, and governance layers; a translational matrix for different stakeholders based on ethical and regulatory requirements in the European Union; and a wireframe, alongside explanatory documentation, for an accountability web-portal to facilitate public participation, transparency, and the work of individuals tasked with human oversight of AI tools. The group consolidated their background research, justification for each resource, and implementation recommendations into a policy playbook.
The second group adapted an existing method deployed in Catalonia by Coboi Lab to the City of Helsinki’s requirements and the specifics of the AI technology. In an extensive playbook, the students present the essential, recommended, and contingent elements of a four-phase participatory process for the introduction of AI technologies. The resource is a detailed, step-by-step guide to integrating participatory and accountability elements into the design, development, and deployment process of public-sector technology.