Developing an Algorithmic Accountability and Data Sovereignty Framework for the governance of Artificial Intelligence platforms in United States education
TL;DR
This study showed how current policy architectures may permit the commercial use of student data and suggests immediate policy changes that would guarantee control of student data, ensure algorithmic accountability, and also prioritize the benefits of students over data monetization.
Abstract
The paper analyzed regulations, standards, and guidelines on the use of artificial intelligence learning platforms in the United States by conducting a systematic review of documents and legal analysis. The data were gathered by purposely sampling 47 documents, including privacy policies, federal and state legislative documents and reported case studies. The analysis identifies recurring themes, including inadequate consent and transparency mechanisms, legal uncertainty across jurisdictions, and documented risks related to algorithmic decision-making. The document analysis suggests that FERPA and COPPA appear insufficiently adapted to address the data processing capabilities of contemporary Artificial Intelligence platforms, which point to meaningful gaps in student privacy protection. Analysis of reported data shows many schools were using Artificial Intelligence with inadequate district oversight, while 60% of principals and 25% of teachers were currently using AI, although only 18% of districts had developed formal guidelines about its usage. Using Data ethics, technological determinism and surveillance capitalism as analytical lenses, this study showed how current policy architectures may permit the commercial use of student data. Reported Case materials suggest risk of algorithmic bias, discriminatory disciplinary actions, and expanded surveillance practice. The study suggests immediate policy changes that would guarantee control of student data, ensure algorithmic accountability, and also prioritize the benefits of students over data monetization.