AI Governance Policies Fragmenting in Higher Ed
As large language models are integrated into university research, publishing, and peer review, institutions are developing fragmented AI policies. This lack of a unified approach risks creating accountability gaps and data management issues across the academic landscape.
The lack of unified AI policies extends to accessibility, creating a confusing landscape for students with disabilities. While some institutions are exploring AI as a powerful assistive technology, inconsistent guidelines mean a tool permitted in one class could be banned in another, creating significant equity and compliance issues. AI-powered tools offer substantial promise for accessibility, with applications like real-time transcription for deaf students, text-to-speech for those with reading difficulties, and voice navigation for users with motor disabilities. These technologies can break down barriers, but only if institutional policies are clear, consistent, and prioritize inclusive design. A major risk of uncoordinated AI adoption is the procurement of inaccessible technology. Without a central governance strategy that includes accessibility criteria, individual departments may adopt AI tools that are not compliant with standards like the Americans with Disabilities Act (ADA), creating new barriers for students and exposing the institution to legal risks. To mitigate these risks, experts recommend involving people with disabilities directly in the AI policymaking process. This ensures that guidelines address real-world accessibility challenges and that any AI tools adopted are designed to be inclusive from the outset, rather than requiring costly and often ineffective retrofitting. The data used to train AI models presents another significant challenge, as biased datasets can lead to inaccessible or discriminatory outputs. A unified governance approach can enforce standards for data privacy and equity, ensuring that AI systems do not perpetuate harmful stereotypes or create algorithmic barriers for students from marginalized groups. Frameworks for unified AI governance are beginning to emerge, offering institutions a structured approach to policy development. Models like the CRAFT (Culture, Rules, Access, Familiarity, and Trust) framework and the AIGN Education AI Governance Framework aim to help universities create coherent, institution-wide strategies that balance innovation with responsibility and inclusion.
