How AI is Reshaping Education: Face-to-Face Interviews for Creative and Critical Thinking

“I am back to college after over six years to pursue my master’s. So far, this is the best subject I have taken as my coursework. Even after having worked on product design and development, I feel I am either new to this world or I never thought in these aspects of product design which I am learning through this course.” Anonymous Student from Mid-term Review, 2025

Transforming Learning Assessment Through “Face-to-Face” Interviews Aided by AI for the AI Era

Are traditional homework assignments and exams still relevant in the age of AI? How can AI support learning without becoming a crutch? In response to these questions, the Product and Process Design class I developed starting in the mid 90's, with a 30-year history and currently capped at 80 students per semester, has replaced traditional assessments with viva voce evaluations. In an era where AI is increasingly embedded in education, instructors face the challenge of distinguishing genuine student understanding from reliance on AI-generated answers. The viva voce method, or oral interview, moves beyond surface-level comprehension by demanding deep engagement and critical thinking. This figure above, contrasts two stages of learning and assessment. On the left, a student uses AI tools alongside traditional methods like reading, assignments and homework's to prepare. It represents the integration of modern and classic approaches for deeper understanding. On the right, the student demonstrates independent critical thinking during a viva voce, relying solely on his knowledge to articulate and connect concepts to the professor. The figure shows the complementarity of AI-supported learning to unaided demonstration of learnt concepts during assessment.

AI allows for personalization on a larger scale while keeping the instructor actively involved. I conduct three interviews per semester with each student, guiding them through the learning process by offering hints and suggestions when necessary. Unlike standardized exams, viva voce provides students the opportunity to express ideas in their own words, articulate reasoning, and apply concepts in new contexts. While AI-driven tools are permitted for homework and assignments, over the past few semesters since AI emerged, I use them solely for feedback rather than grading. An AI tutor, typically a MyGPT with in-context learning, is provided soon after the concepts are introduced, ensuring that students engage deeply with the material before their interviews.

In-context learning refers to a language model's ability to perform tasks based on examples provided within the prompt, without any parameter updates. It allows the model to learn from a few demonstrations given in natural language format, enabling it to address new tasks without fine-tuning. In-context learning is an emergent behavior in large language models where the model uses the provided examples to "locate" latent concepts it has acquired during pre-training, allowing it to generalize and perform the given task. This technique is often employed in few-shot learning scenarios, where a small number of examples are used to guide the model's understanding and execution of a specific task.

To ensure a fair and thorough evaluation, I record the text of these interviews and use AI-assisted templates to assess responses only when feedback is requested by the student. AI has proven effective in helping me give feedback since i perform the same repetitive tasks, by aligning student answers with curated responses. I then read the AI response confirming it is giving the right feedback. Since my answers are elaborated and written with great thought earlier the GPT tends to perform well most of the time for the similar questions I ask the students. I always grade right after the interview and make comments to check against the AI only if needed. My coaching team and I assign grades soon after the interview, and AI-generated feedback has consistently aligned with our assessments. In this manner we are able to scale and maintain a higher level of quality over a larger number of students. Overall the interviews take 8 minutes for 80 students, for each time I interview. While that is a lot of time over the semester, however we have inverted the class and made micro-modules, so I do not lecture at all over the semester. I interview the students, get a hands-on experience into how they are thinking, and spend much more time on their project giving feedback to them.

This shift to viva voce signifies a major transformation in student learning evaluation. It emphasizes not just knowledge acquisition but the ability to reason, connect diverse concepts, and apply knowledge creatively. By encouraging real-time engagement and deeper critical thinking, this approach addresses concerns regarding AI-driven shortcuts in homework-based assessments.

Learning in the Age of AI: The Value of Face-to-Face Interviews in Assessing True Understanding

“I enjoy the class activities that require us to do more than just 'notes and homework.' The activities where we are drawing, storyboarding, solving visually are what truly allow me to understand the lecture but also understand the background steps of product design.” Anonymous Student from Mid-term Review, 2025

Homework help websites such as Chegg, Studypool, Quizlet, and Photomath have changed how many students approach academic challenges, often prioritizing quick solutions over deep comprehension. Platforms like Chegg boast millions of subscribers, while Quizlet and Photomath continue to attract global audiences. This trend raises important questions: Are students engaging in true learning, or is the degree merely a credential on their resumes? Are traditional teaching and assessment methods outdated? With individualized and a large number of questions I ask of the students in different ways for each, often guiding them when they are further away, I do not see the students able to share interviews that easily with others.

Through viva voce interviews, students are encouraged to explain, connect, and reflect on their learning without relying on external aids. This face-to-face assessment enables instructors to gauge not only comprehension but also a student's ability to synthesize information and apply it in meaningful ways.

How Viva Voce Works in the Classroom

“Among all the innovative and engaging activities and sessions planned by the professor, I liked the case study and also the individual interview idea. The case study provides a platform to ponder upon the concepts learned and helps them link with practical scenarios to strengthen the learning. And the interview aspect unveils the shortcomings in the learning effort which is highly possible in the online learning environment.” Anonymous Student from Mid-term Review, 2025

Students in my class face both straightforward and meta-level questions that require them to connect different concepts and frameworks. In their final exam, they are challenged with open-ended problems that demand the creation of unique frameworks to approach solutions. By requiring students to draw connections between seemingly unrelated topics, we assess not just their retention of material but their ability to abstract and apply these concepts to real-world scenarios.

A key benefit of viva voce is the opportunity for immediate feedback. If a student is close to understanding but requires guidance, I offer hints or reframe the question to encourage deeper thinking. This interactive approach results in a more authentic assessment of student learning. Additionally, students may be asked to analyze real-world products or models, allowing for deeper discussions on design choices and architectural principles.

The Role of AI in Modern Learning

Our design class integrates open-ended projects that test a student’s ability to apply learned concepts creatively. AI tools, while useful for generating responses, often fall short in handling complex reasoning or problem-solving. Despite the advancements in AI embeddings, they are still not fully reliable. However, AI plays a crucial role in scaling personalized learning. With increasing student numbers and limited faculty resources, AI supports workflow optimization, enabling more effective feedback mechanisms. Currently, our class is limited to 80 students to maintain quality, but AI allows for greater scalability without compromising student engagement.

Students must learn to use AI responsibly—enhancing their reasoning rather than depending on AI for solutions. By asking interview questions that extend beyond AI-generated responses, we ensure that students can reflect on implications, adapt knowledge to unfamiliar situations, and demonstrate genuine problem-solving skills. Scenario-based assessments within viva voce further validate a student’s ability to synthesize information and apply it effectively. The goal is to cultivate independent thought and adaptive learning strategies in a rapidly evolving educational landscape.

The Generalizability of the Design Interview Model

“The syllabus is structured in such a way that it introduces new concepts and also emphasizes the importance of basic concepts like storytelling and presentation, which at times gets overlooked. This course has a variety of engaging activities, from HBR readings to innovative assignments and drills, helping me immerse in the study.”

While traditionally employed in design education, the viva voce model is proving to be a valuable assessment method across multiple disciplines. Unlike standardized testing, which often prioritizes rote memorization, interview-based assessments encourage students to think critically, articulate their reasoning, and demonstrate deeper engagement with the subject matter. This approach is particularly effective in fields that emphasize subjective analysis and creativity, including humanities, business, healthcare, and STEM education.

By focusing on conceptual application rather than factual recall, face-to-face interviews address the growing need for AI-resistant learning strategies. In humanities and social sciences, this model fosters deeper discussions and critical argumentation. In business and management, case-based discussions and decision-making exercises benefit from interactive assessments. In medical education, problem-solving in patient diagnosis aligns with viva voce techniques used in Objective Structured Clinical Examinations (OSCEs). Even in STEM disciplines, interview-based evaluations assess deeper conceptual understanding beyond what AI-generated answers can provide.

As AI continues to automate factual recall, subjective learning—encompassing creativity, reasoning, synthesis, and conceptual application—will become increasingly valuable across all educational disciplines.

Challenges of AI-Driven Learning

The rapid advancement of AI presents challenges in education, as students increasingly rely on chatbots and machine learning models for assistance. While AI can generate quick responses, it lacks the depth and nuance required for complex critical thinking. Additionally, AI’s probabilistic nature can lead to misleading or incorrect answers, a phenomenon often referred to as AI “hallucination.”

In my course, AI is integrated as a supportive tool rather than a replacement for critical thought. AI-powered tutors provide personalized feedback by aligning content with course material, allowing students to engage in interactive learning. However, viva voce remains the ultimate measure of comprehension, bridging the gap between AI-driven automation and human-driven analytical reasoning. Interview-based assessments not only test factual knowledge but also challenge students to apply concepts in novel scenarios, fostering adaptive problem-solving skills.

Conclusion: A New Frontier in Assessing Learning

“I think he has done a good job of laying out this course for each individual to succeed. The grading scale is interesting, but it works in favor of students that truly put their time and energy into the assignments and enjoy the creativity.” Anonymous Student from Mid-term Review, 2025

As AI continues to reshape education, viva voce interviews offer a powerful method for assessing genuine learning. By emphasizing critical thinking, creativity, and real-time reasoning, this model moves beyond traditional exams and AI-generated solutions. Developing a robust interview-based evaluation framework ensures that students acquire the skills necessary to navigate an AI-driven future—where the ability to learn, think critically, and adapt will be the defining measure of success.

Acknowledgments

I would like to acknowledge the coaches both present and past in the Product Design class at Purdue: Asim Unmesh Ananya Ipsita Duan Runlin Shao-Kang Hsia and many from the past. Their contributions and insights were invaluable in shaping the findings presented in this article. The graphics support is also acknowledged. I also would like to acknowledge the support of Purdue University Online Dimitrios Peroulis and the team in their strong support and belief in past developments that allowed us to scale the class while maintaining a high quality of interaction in its early stages of development. I also would like to acknowledge Frank Incropera for his steadfast belief and understanding of Design and its importance to the U.S. economy. Often the front end of design and creativity is confused with CAD and also manufacturing for production by many academics as well.

Disclaimer

The views, opinions, and interpretations expressed in this article are solely those of the author and any individuals quoted. They do not reflect the official policies, positions, or endorsements of Purdue University or any affiliated institutions. Any mention of research, projects, or affiliations is for informational purposes only and does not imply institutional endorsement.