# Learning Science Lab | Amal Shehata ## **Addressing student AI use with active learning strategies: Amal Shehata, Accounting** If you’re worried about asking your students to go device-free in class, you might be surprised at how receptive they may be to the policy. Through our Faculty Innovation Fellows partnership with Amal Shehata, a Distinguished Clinical Professor of Accounting, we explored how restricting AI use during class and removing restrictions outside of class could address faculty concerns about AI while meeting student needs. ### **The problem** To encourage effective AI use, Professor Shehata had given her students an assignment that prescribed a process of vetting and revising AI output, but it wasn’t working. We turned to the [Barbell Strategy](#barbell-strategy) for student AI use, which recommends having a clear separation between AI-restricted and AI-integrated student work. As part of this strategy, we: 1. Removed devices from the in-class experience. 2. Included short quizzes, in-class activities, and brief current events presentations to track student learning. 3. Removed restrictions on AI use outside of class. ### **The result** After Professor Shehata implemented a no-device policy in her Spring 2026 undergraduate *Principles of Financial Accounting* course, 76.1% of students reported that the policy helped them focus compared with classes where devices were permitted. ### **The takeaway** By using simple strategies such as a no-tech policy or frequent, low-stakes assessment, you can ensure that students are learning foundational knowledge without needing to worry about their AI use outside of the classroom, while creating space to design more intentional, authentic AI-powered assignments. ## **Our journey from the “mushy middle” to a tech-free classroom** ### **Hearing from students: why managing AI use outside of class doesn’t work** Professor Shehata had been thinking critically about what effective student AI use looks like and created an assignment for her undergraduate *Auditing and Blockchain Technology* course intended to give students hands-on experience with AI tools and AI literacy. However, a few students reported to Professor Shehata that because they were already using AI tools in their work, they felt that the assignment design was not authentic to how they used AI for coursework and in their careers. Professor Shehata wanted to better understand her students’ critiques and chose to partner with our team at the LSL. Drawing on our expertise in learning design and experience addressing generative AI in teaching and learning, we designed and facilitated a focus group of six students from the *Auditing and Blockchain Technology* course to gain insight into how to develop a more effective AI-powered assignment. Below: original assignment that didn’t work ![The original assignment that didn't work](/sites/default/files/styles/large_825px_wide/public/2026-10/image4.png.webp?itok=MzA2BXIk) When we asked students in the focus group how faculty could effectively assess their learning, they surprised us by expressing support for restrictions on the use of AI for in-class activities and exams. Rather than focus on improving the original AI-powered assignment, the students discussed how faculty could improve the in-class environment and identified more engaging discussion and graded participation as means to incentivize learning and create a more dynamic classroom environment. ### **Our solution: the Barbell Strategy** Our findings from the student focus group highlighted two different ways faculty could approach AI in teaching and learning: creating AI-free environments to assess foundational knowledge and designing opportunities for students to use AI authentically. We can visualize these approaches using the *barbell strategy*: ![The barbell strategy](/sites/default/files/styles/large_825px_wide/public/2026-10/image3.png.webp?itok=wgv4UWMl) The left weight represents AI-free assessments of foundational knowledge, which students saw as a fair and effective way to measure their learning. The right weight represents assessments where students use AI in ways that are authentic to their studies or future work. Students described examples such as creating AI tutors and using AI to prepare for an interview with an industry expert. The middle represents ineffective AI use, including assignments that prescribe AI practices that feel inauthentic to students. Professor Shehata's assignment had fallen into this “mushy middle.” Rather than redesign the AI-integrated assignment, Professor Shehata decided to remove it and focus on ensuring students could demonstrate foundational knowledge through two strategies: 1. Including short quizzes, in-class activities, and brief current events presentations to track student learning and participation throughout the semester. 2. Establishing a no-tech policy to support active engagement and eliminate in-class AI use. Professor Shehata implemented the new assessment strategies and policy in her Spring 2026 *Principles of Financial Accounting* course. ### **While there was some early resistance, students appreciated the no-device policy over the course of the class.** In a midsemester survey of 46 undergraduate students, 76.1% reported that the no-device policy helped them focus compared to classes where devices were permitted. ![Student survey results showing the no-device policy became more favorable](/sites/default/files/styles/large_825px_wide/public/2026-10/image1_0.png.webp?itok=Rs43RAUC) While students reported mixed feelings about the no-device policy at the beginning of the semester, by midsemester, 65.2% were somewhat favorable or favorable toward it. Students also found the active learning initiatives to be helpful to their learning. Nearly all students found the short quizzes at the beginning of class to be helpful, while the majority of students found brief in-class activities and the no-device policy to be helpful. ![Student survey results showing the course strategies were helpful](/sites/default/files/styles/large_825px_wide/public/2026-10/image2.png.webp?itok=pSFS0doy) Professor Shehata asked an additional qualitative question about how AI can be effectively implemented in Stern classes. Students responded that a clear AI policy, assignments where AI supports rather than replaces student thinking, and device-free environments in class can help promote effective AI use and learning. ## **Conclusion** Professor Shehata demonstrated that even in an AI era, tried-and-true active learning strategies can create an engaging environment where students build and demonstrate their understanding of foundational concepts. She plans to keep using in-class activities and a no-device policy in future courses, while continuing to explore how students can effectively use AI in accounting. Creating intentional boundaries around AI use can give faculty more space to thoughtfully explore where AI can support learning and how to design effective AI-powered assignments. ### **Interested in trying a new tool or pedagogical approach in your classroom?** We are always reviewing requests for collaboration, and we identify our yearlong fellows each summer. Fellows are expected to actively participate in innovation, providing feedback, assisting in data collection, and regularly meeting with the LSL team. To inquire about partnering with us, send us an email at .