Executive Summary 

• Arizona State University’s (ASU) recently launched the first demonstrated proactive and responsible artificial intelligence (AI) rollout in a classroom, highlighting AI’s transformative impact in higher education.  
• Despite the future of AI integrated work likely being handled by university graduates, academic literature and congressional discussion on AI in education has focused almost entirely on students in grade school. 
•  Policymakers should encourage partnerships between US universities and AI industry leaders to create transparent frameworks for how and when to use AI in the classroom, to more broadly encourage responsible and efficient adoption of the technologies. 

Introduction 

After generative AI took the world by storm with the release of ChatGPT in November 2022, institutions of higher education responded by largely limiting or completely banning the use of generative artificial intelligence (generative AI) tools, including ChatGPT, for any work inside or outside the classroom. Rather than embracing the technology that has overtaken work typically completed by a college graduate and has motivated almost half all of employers to cut jobs in 2024 alone, leaders in higher education did what most humans do – stay comfortable in the status quo. Some schools, however, have taken a more proactive approach, such as Arizona State which partnered with OpenAI to harness the ChatGPT enterprise to facilitate student success and innovate academic curriculums for the future. 

Without more programs like ASU’s, higher education risks neglecting the reality of an economic landscape dominated by AI and more importantly failing to harness the power of new technologies to enhance the education experience. A recent survey of 1,100 higher education professionals highlight that burnout, far and away fueled by the need to adapt and use AI, is fueling a growing sentiment for those to leave the field. Asking educators to integrate AI and generative AI tools, without the training to do so, into curriculums that have been largely stagnant in their form for decades often places a significant burden on professors and administrators who are unsure when, how, and to what extent utilize AI, and students being uneasy about the traditional college experience being sufficient to secure a successful career in a world dominated by AI. 

Despite the need for AI guidance in higher education, much of the discussion of AI in the classroom up until now has been centered on the K-12 curriculum, not higher education. And unfortunately, the university presents two main differences that contrast with K-12 learning that make implementing AI more complicated. First, university students often take courses in lecture rooms amongst hundreds of their peers, unlike K-12 courses that typically have smaller classrooms. Second, university students undergo curricula that chiefly, amongst other reasons, is intended to make them productive members of society in a specific field and requires them to learn specific sets of skills rather than a general education. 

The university education of tomorrow and the educators of the present are being left behind by a technology that is moving too fast for any one department or professor to keep up. AI has the capacity to transform how the educational experience is implemented and revolutionize how information is retained and learned, while appropriately preparing graduates for the next evolution of the digital age. However, without frameworks, universal guidance, and collaborative discussion on successful and unsuccessful ventures amongst collegiate leaders, confusion and lack of implementation will continue. ASU’s collaboration with OpenAI, and their transparent challenge inviting faculty and staff to submit proposals that can heighten the positive effects of AI, is a bright light and potential guidepost for future endeavors by peer institutions. Training faculty on how to properly utilize AI, using the technology to ensure students are learning at the intended pace and interjecting applications of it where students fall behind, and curating curriculums to readily adapt to the ever-changing job market landscape are all ideals that should be strived for in this effort of AI in universities. With anything in the AI space, the motivation from leaders at universities should be to be weary of how they intend to implement these benefits to mitigate the potential harmful effects that it may also bring. 

Current AI Training for Educators 

Education thought leaders have argued the need for providing educators with a fundamental framework of what AI is and how to use it to successfully integrate AI in higher education classrooms, but thus far much of the discussion of AI in the classroom up until now has been centered on the K-12 curriculum. These efforts fail to guide higher education institutions for two main reasons. University courses, unlike K-12 courses, often take place in environments of hundreds of students, not dozens; and university education, unlike K-12 learning, is intended to supplement students with sets of specific skills that will make them employable in a certain job market. 

The challenges that AI bring rival the benefits it offers to higher education if professors are unable to understand and properly use the technology. The principal problem to address is “how [institutions and personnel] develop and adapt” to AI. Maybe more so than ever in the age of innovative generative AI tools, using it in the university requires educators not just to look at the surface level capabilities of technology and how it can be implemented, but understand what effects on education each decision to bring in AI has on students.  

Because the focus thus far has been grade school, and the lessons don’t translate well to higher education, using AI in the classroom at universities is not yet widely adopted. But like the wave of professors moving from chalkboard to projector and students from notebook to computer, the move is inevitable. Waiting for the technology to advance further before adopting it only presents a greater learning curve for educators to overcome. Less than a quarter of faculty members used AI tools last fall while almost half of students did, and the vast majority of faculty using them have major limitations in their knowledge of the technology being implemented. Most of those faculty also are using AI only to detect if their students are malevolently using it. Universities agree that knowledge of AI usage will positively impact students’ careers and students’ adoption and willingness to learn about AI tools showcases this. Faculty, however, lack the time and resources to adequately adopt it in their classroom.  

There are some programs being explored, however, to address the issue. Schools such as Stanford, Michigan, and Vanderbilt have begun the creation of advisory boards that bring together students, faculty, and leaders in the AI space to establish guidance and best practices for AI tools. Paired with the establishment of programs and resources lead by technology leaders for educators to learn the technology, this has the chance to lead to success for professors rolling out AI in a just manner. Stanford, Michigan, and Vanderbilt are examples of institutions who have already implemented such boards and groups, in efforts to embrace AI but doing so with a gameplan. Institutions discovering what works best for their own personnel, and working with AI professionals to roll it out, is a situation that leading universities believe will carry AI in the classroom to the finish line. While the results of such approaches have yet to be seen, this wait-and-see thoughtful approach seems promising. 

Incorporating AI To Evolve A Legacy Educational Framework  

Immediate applications for AI into the classroom are being envisioned through two main buckets. The first is to enhance the typical instructor-student dynamic to allow everyone in the classroom to absorb the necessary material rather than some students being left behind in a fast-paced collegiate environment. The second is to build on self-directed methods of learning in a professor’s given curriculum, but also adapt in real-time to specific job markets, by harnessing the power of AI to efficiently target techniques that (most logically for their own careers) to enhance student knowledge. 

First, institutions can use AI to identify gaps in learning and improve educator skills to be sure that everyone in the room is operating at an optimal level. “Learning is a social exercise; interaction and collaboration are at the heart of the learning process.” Much of the promise of generative AI and LLMs lies in their ability to quickly resolve and limit the interaction and collaboration needed in any process, which inherently pivots its most basic use-cases against the standard basis of a university education. A longstanding complaint about the university learning environment is the disconnect between professor and student in courses seating hundreds of bodies at one time that leads to educators not truly knowing how or what any one student is learning until they are assessed at the conclusion of a course. One recent development out of the University of Córdoba, which mimics similar work being done elsewhere, looks to solve this problem by channeling AI to identify how students are learning. The AI algorithm uses data from performance on online quizzes, time spent on online resources, and completion of tasks all located within the professor’s internal website. It then buckets students into one four categories of current readiness in the course and informs university faculty on how they can adapt their learning during the course to best cater instruction where all students may gain sufficient knowledge of a subject. 

Second, self-directed learning can take new forms with AI as faculty and departments may incorporate their own curricula and study materials into tools, to provide a cooperating framework that leverages the power of AI while having a proven and personalized professor in the backdrop. AI can also assist in helping professors and students identify, in real time, the current strengths or skills desired and needed in specific professional industries so that learning in the classroom can build on that needed skillset. Catering learning in courses over hundreds of students to individual learners while changing curricula in real time to adapt to needs of the job market is the AI classroom that many hope and desire, but must only come if universities can implement it with faculty and students both being fully knowledgeable and comfortable in the new environment.  

While there are many potential benefits that could be developed with a collaborative AI framework, policymakers should also consider the challenges to assure student privacy and education are not harmed. The bulk of these challenges are the same ones that AI implementation faces in every industry: access, bias, privacy, and more. The one non-negotiable hurdle that AI integration must overcome in the classroom is striking a balance of its usage so that the interaction and collaboration that formulates an education are not forgotten. How universities tackle this problem remains unknown, but policymakers should begin a patient and collaborative approach from the top-down to make the solution worthwhile. 

Conclusion 

The path from the current university curriculum to one heavily and successfully integrated with generative AI will be a long one but provide significant benefits to students. Coalition between professors, departments, and universities should be formed, and consultation with AI professionals must occur consistently and with dedication for an AI classroom to become the standard of higher education.