Middle School Students as Rights-Based Policy Makers: AI Blueprint Bill of Rights Simulation (Poster 2)

• In Event: Opportunities and Challenges in Teaching Algorithmic Justice in K–12 Computing Education

Abstract

Objectives or purposes: Recently, government entities have laid out frameworks for AI regulation. Notably, the European Union has adopted a risk-based approach, while the United States' Office of Science and Technology Policy has put forth a rights-based approach with the AI Blueprint Bill of Rights. By introducing this rights-based framework to middle school students, we can simultaneously introduce knowledge of AI and fundamental rights of US citizens. This poster showcases the outcomes of a lawmaking simulation conducted by middle school students. During the simulation, these students thoughtfully considered and voted on rights-based provisions for a potential AI bill.
Theoretical framework: We adopt an experiential learning model for students to engage in topics of rights-based lawmaking (Kolb, 2014; Sands & Shelton, 2010). The program was held at the Edward M. Kennedy Institute, which houses a full-scale replica of the U.S. Senate floor, and students acted as Senators to share their opinions and vote on provisions.
Methods and Data Sources: We facilitated a lawmaking simulation with 75 middle school students. Before the simulation, students engaged in a 7-lesson, “AI Bill of Rights” curriculum that addressed how AI might impact US citizen’s rights. In the lawmaking simulation, students were split up into four distinct groups to represent different stakeholder entities: law enforcement, hospitals, schools, and corporations. Within each group, students discussed and voted on provisions tied to the rights of privacy, non-discrimination, and safety. In the last part of the simulation, all four stakeholder groups convened to share their proposals and discuss before the final vote. We report on data collection from the simulation. Notes were taken in each stakeholder group to capture the discussion for each provision and final votes for each.
Results: Notes from the lawmaking sessions showed that students were able to engage in the legislative process and state their opinions in relation to rights. For example, one student in the hospital group shared their opinion on the non-discrimination laws: “Because with [provision] 6 there’s still that chance that the AI might be discriminatory, that’s not stopping the problem. [Provision] 5 is stopping the problem, and [Provision] 6 is just damage control. So 5 is better than 6.” In terms of voting for individual provisions, we found that students tended to vote unanimously, and the highest restrictions in the safety category were voted for in all four of the groups.

Significance: This work encourages a new approach in AI literacy by applying AI knowledge to citizenship and introducing students to their fundamental rights as citizens of the US. Students applied this knowledge to a tangible lawmaking process.

Authors

• Daniella DiPaola, Massachusetts Institute of Technology

• Cynthia L. Breazeal, Massachusetts Institute of Technology