In the summer, Professor G. Sue Kasun of Georgia State University decided to experiment with a new approach for her course on weaving identity and culture into language education. She turned to Gemini, Google’s generative AI tool, asking it to recommend readings, design activities such as writing poems or creating visuals, and even draft grading rubrics. For her, these results served as starting points. She carefully edited and aligned them with her teaching objectives. The outcome proved valuable. She described the process as a significant time-saver.

Professor Kasun reflects a wider movement among faculty who are beginning to treat AI as a collaborator. The approach is not about replacing educators, but about using technology to enhance teaching and save time.

How Widespread Is AI Use Among Educators?

A survey conducted by Tyton Partners with more than 1,800 higher education professionals revealed a steep rise in AI use. Around 40% of administrators and 30% of instructors reported weekly use of generative AI, compared to only 2% and 4%, respectively, in early 2023.

Further insights come from Anthropic, the company behind Claude, which studied around 74,000 conversations with users holding higher-education email addresses over an 11-day period. The analysis classified AI use into distinct categories.

Excellence in Teaching Online

The study found that 57% of AI use focused on curriculum development, including lesson plans, assignments, and interactive simulations. 13% was related to academic research. A smaller share involved administrative tasks such as budgets, recommendation letters, and meeting agendas. About 7% related to grading.

One notable observation was the creative use of AI to design interactive online simulations and web-based games. Faculty often engaged with the technology in a back-and-forth manner, prompting, refining, and adjusting outputs until they aligned with student learning goals. This collaborative style of use highlights how educators prefer to guide AI rather than accept its initial suggestions.

The Grading Dilemma

Although AI shows promise in curriculum design, it remains less trusted in grading. Both Anthropic’s research and faculty surveys reflect a similar pattern. While some professors use AI to streamline portions of assessment, human oversight continues to dominate. Faculty members in a survey conducted with Northeastern University described AI grading as the least effective function of the technology.

Marc Watkins, a lecturer at the University of Mississippi who researches AI in academia, raises a critical concern. He warns of a potential cycle where students rely on AI to generate assignments, while professors depend on AI to grade those same assignments. Such a pattern, he argues, threatens the foundation of education, weakening the human connection and intellectual effort that learning requires.

Professor Kasun shares similar concerns. She supports AI for idea generation and course planning but does not allow it to determine final grades.

Need for Guidelines and Thoughtful Collaboration

Despite the growing use of AI, many educators point to the lack of institutional frameworks. Professor Kasun compares the experience to being alone in the forest, with little structured guidance on how to move forward responsibly.

According to Drew Bent, education lead at Anthropic, universities and technology firms should work together to set balanced norms. He emphasizes that companies should avoid imposing strict directives. Instead, they should encourage collaboration with educators to define best practices that serve students and faculty alike.

Both educators and developers acknowledge that choices made today will influence the future of learning. Decisions about how and where AI fits into the classroom will shape academic culture, teaching methods, and the ethical expectations of future generations.