Project-Based Learning in an AI-Inspired Era
Mark Frydenberg on enriching project-based learning with AI tools
The return to project-based learning, paired with today's AI tools, has created a new learning paradigm. Here, Mark Frydenberg, distinguished lecturer of Computer Information Systems and director of the CIS Sandbox at Bentley University explores the sweet spot between project-based learning — a time-tested teaching and learning practice — and the latest in powerful AI tools.
Anton Spiridonov, a senior at Bentley University majoring in Computer Information Systems, demonstrates his chatbot at a recent technology projects open house. (Photo courtesy Bentley University. With permission.)
Mary Grush: Is there a resurgence of interest in project-based learning? If so, why are we seeing this now?
Mark Frydenberg: Many educators are turning to a project-based learning approach in their courses. Project-based learning puts students in the driver's seat, challenging them to tackle real-world problems and take ownership of their own learning. In doing so, they design and create their own projects and demonstrate that they have mastered course concepts. Project-based learning gives students a story to tell, rather than facts to recall.
Project-based learning gives students a story to tell, rather than facts to recall.
Companies are looking to hire students who have knowledge about a particular subject domain that they studied in college, but they also demand that students have essential skills in communicating effectively, using technology to solve business problems, working in teams, and analyzing information carefully. Giving students opportunities to develop these skills prepares them for the workplaces they are about to enter.
In this era of AI, one of the biggest pitfalls of AI in education is over-reliance. It's tempting to copy and paste a problem description into a generative AI tool and ask it for a solution. While AI tools can help students consider different perspectives and are great for brainstorming, projects are generally more open ended and require students to think critically about multiple aspects of a complex problem.
Grush: Does AI have an effect on the design of project-based learning? Conversely, does project-based learning provide a richer environment for AI-supported programs?
Frydenberg: A challenge for educators today is to design assignments that are difficult for AI to complete fully. That's not always easy. We know that students are using AI tools. One approach to take, is to design projects for students that use AI not to obtain a solution, but as a tool to provide information that they can further analyze. We saw this in the AI hackathon we held shortly after ChatGPT was first released. Students used AI tools to generate trading strategies, and then visualized and refined them using financial analysis software. By adjusting their prompts, they were able to improve their financial models without needing coding expertise.
We also encourage students to use AI to help brainstorm, generating ideas. With AI tools such as Perplexity AI (perplexity.ai), performing a literature review that used to take hours can be done in seconds. That frees students to concentrate more on analyzing the information that AI gathers for them. The caveat, of course, is that while AI tools are getting better, they're not perfect. Students are responsible for checking the references or recommendations from AI to make sure they are valid and free of bias. Students also need the information literacy skills necessary to determine what other sources might not have been provided in an AI-based solution. AI allows students to use their time differently — instead of looking up information, they can verify it and begin to form their own connections between topics.
AI allows students to use their time differently — instead of looking up information, they can verify it and begin to form their own connections between topics.
This inquiry-based approach can lead to student learning that is relevant, motivating, and engaging. AI tools can be key players in project-based learning assignments. Because of their often open-ended nature, project-based learning assignments can provide a context for using AI to explore a variety of aspects of the problem that students might have otherwise missed.
Grush: Does project-based learning open up new types of opportunities for students? Do you have any examples at your own institution — particularly examples of technology-enabled, student-driven or student-run projects?
Frydenberg: Yes! Not all project-based learning is done as part of a class project.
Not all project-based learning is done as part of a class project.
Anton Spiridonov, a tutor in the CIS Sandbox, Bentley's technology social learning space, used his knowledge of large language models (LLMs) to develop a "proof of concept" chatbot that other tutors might use to help facilitate student learning. He uploaded pages from an open source textbook and PowerPoint slides to create an LLM that serves as a reference guide. Coding the chatbot required him to apply the theory he learned in his AI class with the programming skills he developed in his Python class, to create an application for his fellow tutors to use. The project also required independent learning on how to access the OpenAI API (application programming interface) to make use of AI capabilities in his program.
A tutor [in Bentley's CIS Sandbox] used his knowledge of large language models (LLMs) to develop a "proof of concept" chatbot that other tutors might use to help facilitate student learning.
Fine tuning the model was key to Anton's success: "I used PyCharm to develop the code, connected to the OpenAI API, uploaded sample notes, and then added model parameters such as how to answer, how much to say, and what to reject."
While the project was created as a demonstration, Anton says that it could be expanded… but he would need support from the university administration and faculty that tutors could be allowed to use this AI tool to access information that they might use when working with students.
Grush: What about student assessment? Does project-based learning support better assessment strategies, especially where AI is a factor?
Frydenberg: The rise of AI use by both students and faculty has influenced how we assess student learning. In one of my classes, I've moved from written quizzes to "conversational quizzes," where, in 20-to-30 minute conversations on Zoom, students explain their solutions (prepared beforehand) to a problem. While this method doesn't scale easily (I managed to meet with 50 students in a week last semester) and such conversations aren't as thorough as what might be covered in a traditional written exam, within a few minutes I can usually tell which students know what's going on and which are struggling. For those who are struggling, the conversational quiz becomes a learning experience as I guide them to complete the solution they couldn't otherwise figure out on their own.
The rise of AI use by both students and faculty has influenced how we assess student learning.
The main reason for implementing this method of quizzing is to see what students know and can do on their own without AI. Students self-assess their progress ("On a scale of 1 to 10, how much of this were you able to do on your own?") and we compare notes. Usually, their self-assessment and mine are the same or within a point of each other.
Grush: When we reflect on the big picture of 21st century teaching and learning priorities, will project-based learning, especially alongside AI-inspired programs, afford lasting core values?
Frydenberg: Project-based learning will play a role in 21st century teaching and learning, and when combined with AI tools, can deepen the learning experience.
Project-based learning will play a role in 21st century teaching and learning, and when combined with AI tools, can deepen the learning experience.
For 25 years now, educators and employers have been talking about the need for 21st century skills (collaboration, communication, creativity, critical thinking), which are essential for students to succeed in the business world. We might also add information literacy (being able to determine what's real from what's not reliable online), technology literacy (being able to use technology effectively to solve problems), and ethical awareness to ensure that Internet-based and AI-based tools are used equitably and responsibly by today's digitally-informed students.
Project-based learning encourages students, instead of memorizing facts, to interact with real-world problems, engage curiosity, and think creatively on solving problems both inside and outside the classroom.
By leveraging AI tools, project-based learning has the potential to transform education by changing students' roles from passive consumers of information to active "creative creators" (to quote New York Times columnist Thomas Friedman). Today's world demands that students demonstrate their abilities to adapt, innovate, learn, and generate ideas when faced with creating new solutions.
[Editor's note: Photo courtesy Bentley University. With permission.]