Chris Rios
Associate Dean for Enrollment Management
AI is transforming higher education. As it does so, universities are racing to provide guidance for students about when and how it should be used. Whether students will use AI isn’t the question. It’s increasingly difficult to do anything online without encountering its integration into the tools most of us use every day. The question is how they will use it—intentionally or passively, responsibly or irresponsibly, well or poorly. As an example of the administrative efficiencies AI can offer, we in the Graduate School are increasingly employing it to facilitate daily tasks both big and small. We use it to help analyze data, draft summaries, and produce reports. We developed AI agents to assist with projects such as customer service and market research. And we work with colleagues to explore ways AI can enhance our students’ learning. The Graduate School can thus serve as a microcosm of the broader landscape of graduate education, one where questions and challenges emerge at a pace much faster than many of us expected.
This past November, NORC at the University of Chicago hosted a convening on AI in graduate education. The event offered a day and a half of robust, thought-provoking conversations about the challenges and opportunities brought by generative AI. Topics included issues of fairness and bias, policies and guidelines, operations and communication, and drivers and influencers. An official report is forthcoming that will include a proposed set of guiding principles to help graduate schools navigate the road ahead. As we await the final product, we would be well served in considering which priorities should guide the integration of AI in graduate education at Baylor. Let me suggest four.
First, graduate education must account for the ethical challenges of accessibility and bias, topics that received considerable attention at the NORC meeting. Accessibility relates to how availability can privilege some individuals and groups over others. If AI has the potential to enhance our work, then those with access clearly have an advantage over those without. At a national or global level this is often described as the digital divide, a phrase meant to highlight the way AI will likely widen the economic gap between rich and poor nations. Yet a similar kind of divide is possible between students. If those with more resources can afford to buy access to the best tools, those without are at a clear disadvantage. Thus providing a baseline of access for all students should be a priority. Baylor is fortunate in this regard. Thanks to our ITS leadership, advanced large language models (LLMs) are available to all students, faculty, and staff. These tools not only help level the digital playing field. Because Baylor’s contracts include robust restrictions about data protection, the systems also safeguard user privacy and intellectual property.
Considerations of bias within AI are especially important when AI tools are used to assist in decision making. Thanks to AI we can process written information, such as teaching evaluations and exit surveys, much faster than before. Should it be used to help evaluate student applications? Should faculty use it to help with grading? In these areas, bias can quickly become an issue. Biases, of course, are not inherently bad, and they are not something that can or should be eliminated. They are (in the language of design thinking) a gravity problem, something that can’t be eliminated, only accounted for. The challenge is to learn how we can use AI in a way that avoids some biases while reinforcing others.
Second, we should commit to transparency about when and how AI is used. This seems obvious when it comes to students, for whom academic integrity and intellectual development remain central concerns. But it’s also important for faculty, staff, and administrators, especially regarding expectations and policies about student use. And it is especially important at this stage, when there is so much uncertainty about what AI can do and what is or isn’t beneficial. This is why the Graduate School asked all programs to create guidelines about the appropriate and ethical use of AI for students in their disciplines. These guidelines are intended to not only help prevent overreliance on the tools as students acquire the necessary knowledge and skills but also allow them to learn appropriate uses that are becoming normative in their fields.
Third, and related, faculty and students should be aware of emerging disciplinary and professional expectations about policies and best practices. Some disciplines are rapidly embracing AI technology and are preparing students for careers that already expect proficiency. Others are taking a slower approach. But virtually all are finding explicit guidance necessary. For example, there are emerging policies about ethical and appropriate AI use from funding agencies and publishers.
The pace at which these needs have arisen puts many faculty in a difficult place. Guidance can only be provided by those who know what LLMs can and can’t do, but many faculty lack this knowledge. Some have no interest. Some believe the learning curve is too steep. This situation should motivate educators to take the time to learn for themselves for the benefit of their students.
Finally, AI use must be oriented towards the flourishing of our students, faculty, and staff.
One common and important concern about AI is how it will affect activities and processes at the heart of our mission as educators. Will overreliance cause students to fail to develop core cognitive abilities? Another common concern is how AI has the potential to put people out of work or reduce the sense of value they offer. Since financial stability and a sense of meaning and purpose are key components of human flourishing, these concerns must be considered as AI becomes integrated into our personal and professional lives. If AI doesn’t make students better scholars, researchers, and professionals, if it doesn’t make people feel more valuable and purposeful in their work, then AI use should be stopped or redesigned so that it contributes to rather than hinders human flourishing.
In closing, let me offer two brief stories that have framed my perspective about the challenges and opportunities ahead, one from the world of math and one from art.
In the first half of the 20th century, the highest-level math taught at most US high schools was algebra. By the end of the century, it was calculus. What accounted for this change? One key factor was the invention of the scientific calculator in the 1950s. Significantly, the calculator didn’t diminish students’ ability to learn math as much as it allowed them to pursue higher levels within the same amount of time. Imagine what the next generation of students will be able to accomplish within the years of their undergraduate and graduate degrees.
The camera was invented in the first quarter of the 19th century. Within a few decades, photographs were becoming common and artists were experiencing something akin to an existential crisis. What was their purpose now that a machine could capture reality better than they could? What was the next major artistic school to emerge? Impressionism, the style that captured human perception of everyday life, not by striving for increased realism but by violating long-established artistic norms.
These stories lead me to hopeful conclusions about the way we can respond to the challenges and opportunities brought by AI. These technologies will touch virtually every area of modern society and have profound influence on our life and work. They will have the potential to enhance or diminish activities often taken for granted. The question, then, is not whether AI will replace human creativity and agency. The question is whether Baylor graduates will develop the wisdom and courage needed to shape their influence—to lead and serve in ways that contribute to human flourishing and thereby help do God’s work in Creation. Our graduate students deserve our best effort to prepare them for what lies ahead.