At UC Berkeley’s Homecoming event on October 9, 2026, John DeNero, a teaching professor in the Department of Electrical Engineering and Computer Sciences, shared insights into the integration of artificial intelligence (AI) in computing education. DeNero, who is also the faculty director for the Data Science Undergraduate Studies program, was awarded the Fiat Lux Faculty Award earlier this year for his contributions to educational innovation, including the establishment of the university’s data science undergraduate major.
During his address, DeNero highlighted both the challenges and opportunities presented by AI in education. While acknowledging that the use of AI tools is not without its drawbacks, he emphasized UC Berkeley’s efforts to enhance educational outcomes. “The reality of how AI is being used in education is not all positive at the moment,” DeNero remarked. “Then I’ll walk you through how I think Berkeley is leading the way in making it a lot better.”
DeNero referenced recent claims by OpenAI regarding the solving of numerous math problems, prompting a reflection among educators about the role of human expertise in technical fields. “The rate of progress and the problems that people are working on is just changing really fast,” he noted. He argued that while AI has advanced, the importance of human understanding in computing remains critical. According to him, a 2026 study by Anthropic revealed that AI success rates varied significantly, from 39% among novice users to 65% for experts, underscoring the value of human competence.
“Students who have built software themselves tend to focus on scrutinizing the AI’s output and making specific requests for adjustments,” DeNero explained. He pointed out the necessity of human evaluation in AI, stating, “Every company building AI is trying to understand whether some AI tool is actually solving the problem it’s supposed to solve, and that’s very much a human activity.”
DeNero also discussed effective strategies for teaching computer science, emphasizing the importance of hands-on practice. “Homework is essential to practice applying concepts learned in class,” he said. He argued that merely relying on AI to explain homework answers does not equate to real learning. “It feels productive, but I don’t think it helps them discover how things apply,” DeNero stated.
He offered examples of AI-enabled tools developed in 2023 to support learning without giving away answers. Tools like a personalized homework tutor, the 61A Bot, have provided students with constructive feedback on their errors rather than just telling them what was wrong. “The AI system described why their answer was wrong, not just telling them what was wrong,” DeNero explained.
At the conclusion of the talk, audience members, including electrical and computer engineering student Max Nguyen, expressed their gratitude for DeNero’s insights. Nguyen shared his positive experiences using AI tutoring methods in his applied math class, noting that it has significantly improved his understanding of the material. Following this, DeNero responded enthusiastically, celebrating the impact of these educational innovations.
UC Berkeley continues to explore and develop AI tools aimed at enhancing the learning experience for students, with various initiatives underway to integrate human expertise with advanced technology in educational settings.
