The rapid growth of Artificial Intelligence (AI) is creating both excitement and angst on where and how to begin safe, effective integration. While there are many ways AI can be applied, these suggestions are focused on educational delivery. For those who have begun applying and engaging student-facing AI, the vertical evolution requires structured and continuous development of institutional-level AI governance to maintain safe and ethical use. Governance structures should consider continuous growth of policies, guidelines and directives for use, approval processes, and staged training for faculty, staff, and students based on the speed of changes occurring in the market. 

Lacking or vague policies, structure, or training approaches may leave administrators, faculty, and students without the guidance needed to reap the benefits of AI use. Lacking governance structures or broad policies may unintentionally promote unsafe or unethical use of AI on the part of students, faculty, staff, and instructional designers who may not have the proper guidance to integrate. This can lead to compromised course design and program content, AI hallucinations, misinformation, and privacy breaches to proprietary university content, impacting the university and quality of education. 

On 21 February 2023, ACAO hosted a panel discussion on Chat Generative Pre-trained Transformer (ChatGPT) and its potential uses (or abuses) in higher education. The goal of the panel was twofold: (1) to introduce provosts and other attendees to ChatGPT (developed by OpenAI) and similar AI language models and technologies; and (2) to discuss the opportunities and challenges of this new technology on our working, teaching, and learning environments.

Three speakers addressed attendees. First, Dr. Jing Peng from Montclair State University provided a high-level presentation on AI chatbot technology and its evolution over time. Dr. Peng’s presentation highlighted the fact that this technology is not “new,” per se, with roots that extend back decades – but its development is accelerating exponentially as it becomes “smarter” given the vast amount of data that are now available for its refinement and evolution.