Students are back in school which means so are ChatGPT and other generative AI tools. As the academic year begins, students, faculty, and administrators are grappling with the challenges and opportunities presented by generative AI.
Over the summer, we organized a series of workshops with individuals from diverse higher educationinstitutions throughout Oregon, including public and private universities as well as community colleges. Our aim was to enlighten instructors about both the potential benefits and challenges of generative AI for themselves and their students.
We found important common themes across institutions and roles.
Everyone agrees that generative AI has the potential for undermining the development of critical thinking skills. 75% of people rated this as their number one concern, followed by the potential for students to over-rely on the technology and for the widening of the digital divide (58% in both cases).
Generative AI is making waves across various occupational sectors, but its ripple effect is particularly pronounced in the realm of higher education. Higher education roles are significantly overrepresented when it comes to the influence of generative AI. In one important study, out of the a list of the 30 occupations expected to be most affected by generative AI, a staggering 70% are directly tied to postsecondary education, ranging from "English language and literature teachers" to "psychology teachers" and even "education administrators." In contrast, the remaining 30% encompass diverse professions such as "telemarketers" and "procurement clerks."
This concentration signifies that while various sectors feel the touch of AI, higher education is standing at a critical juncture of transformative impact.
Presidents and provosts were more bullish on the opportunities presented by generative AI. About 75% of this group believed that the technology could notably enhance student learning, faculty teaching, and administrative processes. However, community college staff and faculty, who are closer to the frontline, are more skeptical, with only 40 to 45% feeling that these areas will benefit.
A uniform preparedness rating of a mere 2 out of 5 across student, faculty, and administrative groups paints a concerning picture of the current state of readiness. This consistency in unpreparedness not only highlights a significant gap in knowledge and skill acquisition but also underscores a missed opportunity to leverage the collective potential.
The urgency to address this cannot be overstated. The AI that is in use this fall will be the worst AI we’ll know. Institutions and organizations should prioritize comprehensive training programs, allocate resources for skill development, and foster an environment of shared, continuous learning to bridge this preparedness gap and ensure that all groups are equipped to meet future challenges head-on.
A promising approach is to recognize that students are often more skilled and facile with these tools than are instructors. Learning communities and forums which allow the students to become the teacher are worthy of consideration. As one professor noted, “faculty are trying to figure out how to saddle the horses while the students are building a car.”
When it comes to plans to use AI, our survey indicated a mixed sentiment more generally. When asked about their experience or intentions to use generative AI in their work, around 39% of respondents affirmed they have used or plan to use it, while 32% remained uncertain, and 29% indicated no such plans.
The results took an intriguing turn when participants were queried about discussing generative AI with their students. A notable 50% have already engaged or plan to engage in such conversations, underscoring the significance of the topic in educational discourse. However, there's an even split among the remaining participants, with 25% being uncertain and another 25% having no plans to discuss.
It is not sustainable to not discuss generative AI. In surveys of college students, it’s clear that its use is widespread and undetectable. Additionally, students want to learn how to use the tools so that they can be more competitive students and future professionals.
In a separate study, a significant 50% of students expressed their intention to utilize AI tools regardless of guidance or advisories from educators, underscoring the inherent appeal and perceived utility of these technologies. Meanwhile, 23% acknowledged having already used AI for tasks such as exams or assignments. However, a striking 54% shared that their instructors had not engaged them in any dialogue regarding the use of AI tools in their studies, something that students want. Students don't want to be caught by surprise and appreciate when instructors are clear about the use of the tools—providing clear guidance on citation requirements (eg, show your prompts) and offering perspectives on which tools are preferred and why.
Detection is futile. While 96% of human-written content can be identified as such it’s questionable whether this is good enough. Even a 4% false positive rate is too high, especially when people whose first language is not English are twice as likely to be flagged as cheaters.
AI-generated text sneaks past our best tools with a 74% detection rate. More concerning is when AI text is refined either by human hands or another AI iteration: detection plummets to 42% and a mere 26%, respectively. In essence, the technology's sophistication outpaces our capacity to discern its handiwork.
These findings suggest that there is a significant communication gap between educators and students concerning use of AI. Given students' evident inclination to adopt AI tools, it becomes imperative for educators to proactively initiate conversations, setting clear guidelines while also addressing the potential advantages and pitfalls.
As technology continues to evolve, institutions are presented with opportunities to redefine and enhance the educational experience. In the workshops we took a closer look at the transformative impact of Gen AI across various facets of higher education. Here’s where we found the most agreement:
A fundamental shift towards "Dynamic Approaches" is poised to revolutionize teaching. Gen AI facilitates customizable learning pathways, allowing instructors to tailor course materials to suit individual student needs, ensuring more effective content delivery. Moreover, the potential to introduce new curricula, continuously updated and aligned with evolving industry standards and global events, promises a more relevant and contemporary educational experience.
"Adaptive Feedback" stands as the cornerstone for the future of learning with Gen AI. Traditional learning environments can transform into personalized experiences, with adaptive, intelligent partners guiding students through their academic journey. Beyond standard feedback, learners will have access to novel compositions for self-expression, enriching their ability to understand, interpret, and engage with the curriculum in creative ways.
The focus on "Enhanced Productivity and Service Quality" underscores the transformative potential of Gen AI in higher education administration. Enhanced career development resources can offer students a clear path to professional success post-graduation. Simultaneously, administrative tasks can be streamlined, leading to heightened efficiency and improved student support and access. As a result, institutions can ensure a more holistic and supportive academic environment for all stakeholders.
As the academic year begins, it's essential for educators to address the role of artificial intelligence in the learning environment. Chris Dede from the Harvard Graduate School of Education notes, "Anything that generative AI can do on a test, it’s going to do in the workplace." This underscores the importance of preparing students for AI's growing role.
Lack of understanding of the tools is problematic. As one college instructor pointed out, "If you don’t teach the tools, people think they are perfect." It's a direct call for educators to ensure students are both aware and critical of AI's capabilities and limitations.
Stanford's Human Centered AI group suggests educators probe deeper: "Ask your students what they dread about AI." This promotes an open dialogue, allowing educators to address concerns directly. Meanwhile, a community college instructor emphasizes personal contribution: "Let them know you’re far more interested in what they have to say than what ChatGPT has to say."
It's clear that the conversation around AI in education is multifaceted. Educators need to initiate these dialogues early, ensuring students are equipped and informed for the future.
Interested in talking about how generative AI may impact your institution? Reach out: email@example.com.