AI as valuable tool in comparing candidates’ plans for UPR

Artificial intelligence may be an essential tool in helping select the leader best equipped to guide the University of Puerto Rico forward at a critical point in its history. (Erika P. Rodríguez/The New York Times)

By Jean García Rosa

Special to The Star

For the first time in recent history, the evaluation of presidential candidates at the University of Puerto Rico (UPR) has been supported by artificial intelligence (AI) tools. This approach provides an objective, structured, and fair analysis of the candidates’ work plans. While this technology does not replace human judgment, it complements it, facilitating informed decisions in complex processes.

Certification 91 (2024–2025) from the UPR governing board outlines a rubric composed of 39 criteria, covering areas such as Academic-Administrative Leadership, Interpersonal Skills and Conflict Management, Financial Administration, and Student Affairs. Systematically applying this rubric requires a rigorous evaluation, which AI can significantly streamline.

An analysis of the work plans from the eight finalists, using platforms such as ChatGPT (OpenAI), Gemini (Google), Copilot (Microsoft), and DeepSeek, revealed marked differences in candidates’ compliance with the criteria. Candidates Mayra Olavarría Cruz and Héctor Ríos Maury received the highest scores, showcasing solid, structured proposals fully aligned with the institution’s objectives. Notably, Ríos Maury achieved 100% compliance, presenting a plan with 32 strategic vectors, timelines, and performance indicators. Olavarría Cruz followed closely, achieving 96.52% due to her humanistic and practical approach that addressed multiple aspects of the rubric.

In contrast, candidates Edwin Barea Rodríguez, Angélica Varela Llavona, Zayira Jordán Conde and Wilma Santiago earned scores of 78.24%, 75.32%, 68.53% and 62.49%, respectively. Walleska de Jesús Bonilla and Dr. Agustín Rullán Toro scored 61.34% and 60.57%, respectively. While diversity of perspectives is valuable, the lower quality of some proposals may indicate a lack of knowledge about the university environment or an underestimation of the presidential role. Furthermore, the brevity of certain proposals limits their operational viability and complicates evaluation against the established criteria.

Analysis methodology

The evaluation was based on the official rubric, employing a scoring scale from 0 to 4 points per criterion. These scores were then transformed into compliance percentages, allowing for a standardized comparison of each plan’s response. To enhance validity, the results across several AI platforms were compared. The consistency between ChatGPT, Gemini, Copilot, and DeepSeek bolstered the reliability of the analysis, identifying the proposals of Ríos Maury and Olavarría Cruz as superior.

This exercise demonstrates that by applying clear criteria and advanced technology, more objective decisions can be made. Given the significant fiscal, demographic and academic challenges facing UPR, there is no room for improvisation. The university needs leadership that combines strategic vision, executive capacity, and a deep understanding of the system. If artificial intelligence can help illuminate that path, then its use is not only beneficial but essential.

Jean García Rosa, CPA, CFE, CICA, MBA, is a professor at Inter-American University of Puerto Rico, with prior experience as a professor at the University of Puerto Rico.