Embracing AI Era: CMU’s Emphasis on Critical Thinking

Carnegie Mellon University Pioneers AI Integration in Education

At Carnegie Mellon University (CMU), the integration of artificial intelligence (AI) into the curriculum is reshaping the educational landscape. With a focus on empowering students to not only use AI but also to contribute to its development, CMU is fostering an environment of innovation and critical thinking.

Redefining Educational Approaches

Tom Cortina, associate dean for undergraduate programs at CMU’s School of Computer Science, emphasizes the importance of equipping students with the skills to develop the next generation of AI tools. “At CMU, we train students how to be contributors to the next generation of tools,” Cortina said. The goal is to adjust teaching methods to encourage problem-solving and deeper thinking as AI becomes a more integral part of academia.

According to Zico Kolter, associate professor of computer science and director of the Machine Learning Department, learning is most effective when students struggle with concepts and work through problems independently. “This is a transformational technology,” Kolter said, highlighting the need for new teaching approaches in a world where AI can solve many homework problems.

Faculty-Led Initiatives and Best Practices

To address the challenges of teaching with generative AI, SCS Dean Martial Hebert tasked Cortina and Kolter with organizing a faculty summit. The event allowed faculty to share techniques and strategies for integrating AI into education. “We weren’t expecting to have any particular answers per se, but the idea was to share some of the techniques that faculty were trying,” Cortina said.

Tom Mitchell, a Founders University Professor in the Machine Learning Department, discussed the rapid adoption of generative AI and its impact on daily life. The summit encouraged experimentation and adaptation in teaching methods to keep pace with technological advancements.

Incorporating AI into coursework requires a shift in focus. Instead of extensive code writing, students need to understand how code functions and engage in higher-level tasks like software planning and design. “We have to be a bit more agile,” Cortina said, emphasizing the importance of maintaining student ownership of their projects, even when AI assists with coding.

Exploring Broader Concepts Beyond Coding

Assistant teaching professor Mike Taylor introduced “algorithmic thinking” into his “Effective Coding With AI” course, prioritizing problem-solving over the end result. The course served as a “sandbox” for students to explore AI’s potential in a safe environment, leading to a diverse array of projects.

Taylor’s findings, submitted to the Technical Symposium on Computer Science Education, revealed that students who actively engaged with AI rather than relying on it learned the most. The course emphasized ethics and inspired students to consider technology’s societal impact, fostering collaboration and creativity.

Campus-Wide AI Research Initiatives

The Eberly Center for Teaching Excellence and Educational Innovation launched the Generative Artificial Intelligence Teaching as Research (GAITAR) Initiative in 2023. The program supports instructor-led innovations and educational research, focusing on the effects of AI tools on student learning.

Chris McComb, a professor of mechanical engineering and director of the Human + AI Design Initiative, is one of the 27 GAITAR Fellows. His research involves using AI chatbots in coursework to challenge students by introducing deliberate mistakes for them to identify. This approach has led to significant improvements in student performance.

Chad Hershock, executive director of the Eberly Center, noted that the impact of generative AI on learning depends on how it’s used. When AI serves as a thought partner, it can enhance learning by creating opportunities for critical thinking.

Innovative Teaching Practices

Anand Rao, distinguished service professor at Heinz College of Information Systems and Public Policy, encourages students to engage in “meta-thinking” in his “Responsible AI” course. This approach involves framing and deconstructing problems before seeking AI assistance, fostering a collaborative brainstorming process.

In the “Ethics, Safety and Social Impact in NLP and LLMs” course, Maarten Sap, assistant professor at the Language Technologies Institute, gives students vague instructions to encourage independent thinking. This method requires students to operationalize prompts, forcing them to engage more deeply with the material.