Skip to main content
Usable Knowledge

Tips for Using AI, From Grad Students and Professors

A new guide offers advice and strategies for using generative AI tools to support self-directed project-based learning
Logo from Creative Computing toolkit

Powerful generative artificial intelligence (GenAI) tools that can rapidly generate human-like responses to prompts and questions have created a challenge for some educators. Concerns abound about how easy it can be for students to use freely available and sophisticated chatbots to get quick answers for assignments and even churn out entire essays, undermining the learning process. The angst caused by the new tools reminds Professor Karen Brennan, director of the Creative Computing Lab at the Harvard Graduate School of Education, of the “moral panic” that ensued following the invention of the pocket calculator and has led to “a reckoning,” she says, in education.

“In K–12 and in higher education, some of the anxiety around generative AI and potential misuse has been around a realization that maybe we are asking learners to engage in work that isn’t actually particularly powerful or meaningful. If a machine could do it, what does it mean to be asking students of all ages to be doing it,” Brennan explained in a recent interview.

A newly published guide, co-authored by Brennan, Paulina Haduong, Avantika Kolluru, Sally Yao, and Jacob Wolf, demonstrates the productive ways in which GenAI can be used to aid teaching and learning in the context of supporting creative self-directed projects that prioritize student autonomy, critical thinking, and problem solving.  The guide highlights both the opportunities and the difficulties that the new technology presents and provides advice, strategies, and some guardrails to keep in mind when using it. 

To develop the guidance, the research team interviewed 27 HGSE students about their experiences of experimenting with GenAI while working on individual learning design projects as part of the Learning Design, Innovation, and Technology (LDIT) master’s program, and seven faculty members about their students’ work.

General advice:

The guide recommends that educators and students of all ages, in any learning environment that supports self-directed projects:

1) Think about the wider implications of using GenAI.

Be thoughtful and intentional about how you use new AI tools, the guide suggests, keeping in mind their “limitations and potential harms.” Some of the concerns that are highlighted with using the technology are its “tendency to hallucinate, substantial environmental footprint, accessibility barriers due to cost, potential homogenization of culture, and inherent algorithmic biases.”

2) Protect learning and “authentic voice.”

Students are encouraged to think of “GenAI as a support, rather than a replacement, for personal thought, effort, and style.”

Haduong, Ph.D.'23, a visiting scholar affiliated with Brennan’s lab, said in an interview they appreciated how the students in the study used AI tools in their projects as part of their learning process, but not as a replacement for it, and how learners prioritized their own voice being present in their work. One student, for example, advised other students to, “really think of what you want at this moment. Do you just want to get the job done, or do you want to learn?”

GenAI helped learners with what Brennan described as “cognitive offload” in their learning design projects, which ranged from building apps and websites to in-person interactive experiences and curriculum design. This enabled students to focus on the parts of their projects that were the most interesting to them and allowed them to “do more, go further, go deeper,” she said.

3) Play and experiment.

The HGSE faculty and students interviewed for the guide recommended exploring and playing with different GenAI tools. “You’ve got to play around, get a sense of things. You can’t help students until you’ve got a feel for it yourself,” said one faculty member who was quoted in the guide.

Some students found conversational AI tools to be great thought and feedback partners while others used tools like ChatGPT to play “devil’s advocate” and provide opposite perspectives from their own which enabled them to test the strength of their ideas. 

4) Educators: explain your “why” and take a humble approach to learning about new technologies alongside your students.

Instead of blanket polices banning the use of GenAI in learning, Haduong encouraged educators to take a more nuanced approach by getting to grips with the tools themselves and carefully considering when they might be useful and when they might not be. Haduong added that some students, “found it really helpful when faculty explained why they were doing an assignment” because it helped them to “make decisions about what they might want to offload [to the AI tool] and what they wanted to do themselves.”

Learners also valued faculty who took a humble approach in class toward the new technologies, saying something along the lines of: “I'm learning alongside you and the tools are always changing. We don't know everything, but we're figuring it out together,” Haduong added. 

5) If at first you don’t succeed, try, try, again, but only for so long.

Students reported having to make “multiple attempts and prompt refinements” when using GenAI tools. They also warned about the steep learning curve involved and acknowledged that sometimes drawing on their own skills and capabilities was more effective. One student who had a protracted experience using GenAI advised, “If you’re finding that you’re going for six hours, and maybe you should have been done by now, maybe you should stop.”

Usable Knowledge

Connecting education research to practice — with timely insights for educators, families, and communities

Related Articles