Teaching

AI gives higher education a generational opportunity to refocus on teaching students how to learn, not just what to learn. Research shows that students learn most when they actively build their own understanding rather than passively receive it. Emerging technologies, including interactive media and AI, can support active learning by expanding what students can investigate, or undermine it by doing the reasoning and communications for them. These technologies lower the barrier for students to push the frontier of human knowledge themselves. The educator's role then becomes that of a coach: helping students recognize what remains unknown and pointing them toward the tools, methods, and resources to discover and communicate new knowledge.

This shift also requires rethinking what the products of learning ought to be and how we measure student success. I help students use emerging technologies to make discoveries driven by their own interests, so they are intrinsically motivated to do the reasoning and communication themselves. Because students bring different curiosities, cultures, and knowledge into the classroom, I design learning objectives and encourage them to build research questions around the narratives they know and care about, centering perspectives a standard syllabus would miss and giving students who rarely see themselves in course material a direct stake in the learning process. Rather than writing standalone papers and answering multiple-choice tests, my students build research portfolios that document their process of learning and discovery, from pen-and-paper reflections on course materials, to organizing multimedia content around answering an open question, to designing a plan for executing a multi-step project. Rather than having students work in isolation, I apply my research on human-AI group work to create in-class learning assignments that help them collaborate more effectively with one another alongside cutting-edge AI tools.

My teaching philosophy grows out of my experience as an undergraduate research assistant, where data science methods first gave me the tools to develop my own research program, and it now extends that program: I view learning as an expression of narrative agency, in which students integrate new information with their prior knowledge, communicate their understanding to others, and update their beliefs as they encounter new evidence or disagreements. By the end of my courses and mentorship, students can (1) identify gaps in their knowledge and frame their curiosity as a concise, open question; (2) actively verify information and find the best tools and methods to answer that question, including writing code, even with no programming background; and (3) clearly communicate their knowledge so others can learn from and build on their hard work.

Emerging technologies as tools for creative discovery

My teaching and mentorship center on using emerging technologies to expand what students can investigate, build, and communicate, regardless of their technical background. Drawing on ongoing conversations with my mentees about how they use conversational AI in research and coursework, I label tasks as either reflection (no AI) or implementation (AI encouraged). Students can use AI to identify gaps in their knowledge, expand their sources, and find weaknesses in their arguments. They must then verify that information, integrate it with what they already know, and argue their conclusions in their own words, refining them through conversations with their peers and with me. Even when AI writes their code, they must explain how it works, and I am actively developing Socratic dialogue methods that quickly identify gaps in their working knowledge of analyses, arguments, and software, for example by asking students to explain a function's inputs and outputs rather than simply what the function does.

Because AI can produce a finished product, I also assess the process behind the product through research portfolios, whether the final product is a working experiment, a quantitative analysis, an art project, or a conference talk. I evaluate each portfolio with a rubric based on the revised Bloom's taxonomy. The rubric tracks how students progress from applying methods, to analyzing data, to evaluating tools and sources, and finally to creating an original contribution, and each level maps onto the three learning goals described above. I also assess learning through one-on-one and small-group dialogue, where students explain their thinking to me and to one another in a low-stakes setting, so we can evaluate together how each student integrates their background knowledge into how they communicate. In these conversations, I guide students to clarify their language using explanatory virtues studied in cognitive psychology, such as simplicity, scope, and coherence, making their explanations both more precise and more accessible.

Classroom teaching

I have put this philosophy into practice in the classroom. As a graduate TA for Signal Detection Theory and the Cognitive Psychology Laboratory at UCLA, where many students had never written code, I built Jupyter-based labs that taught Python from the ground up, culminating in each student working in small teams to program and run their own psychophysics experiment in PsychoPy. One student wrote: "Even though I have never taken a Python class before, Hunter's teaching made learning Python easy and made me feel confident in applying it to real-world situations."

Beyond in-class instruction in these labs, I also held open office hours for undergraduates to discuss difficult course concepts, such as working through mathematics problems in Signal Detection Theory or discussing research findings on human learning and behavior. As a TA for Dr. Keith Holyoak's graduate seminar, Thinking, I held open office hours to walk through complex analyses in research papers and discuss open questions in the field with curious students. My median rating across all courses is 9/9, and multiple students describe my help as essential to their success in their course evaluations.

Outside my TA roles, I have given guest lectures on emerging technologies, psychology, and communication. For example, in Dr. Zachary Horne's Big Data Psychology course at the University of Edinburgh, I presented the mechanics of the data science pipelines I use to study language dynamics and persuasion on Reddit forums. At UCLA, I lectured on online conspiracy narratives in Dr. Holyoak's Thinking seminar, discussing current experiments and communication studies on how algorithms in online networks reinforce misinformation and conspiracy narratives.

Project-based research mentoring

I have had rewarding experiences mentoring many undergraduate students. Each student develops an original research question, and I coach them through the full research process: from framing the question, to collecting and analyzing data, to communicating their findings in written papers and verbal presentations. Each project becomes the core of the student's research portfolio, which many leverage in the next chapters of their careers. This is because the projects grew directly out of students' own interests, and I trained them in cutting-edge methodologies to advance those interests further. Several students presented their findings at conferences as undergraduates and are pursuing or have completed PhDs at Princeton, Carnegie Mellon, Columbia, and the University of Illinois Urbana-Champaign (UIUC). Below, I highlight four of their successes.

Michelle Bak was a UCLA digital humanities student interested in how health and wellness Instagram influencers portrayed body image to their followers. I trained her to mine 285,000 Instagram posts from top-ranked health influencers and fit embedding-based topic models to identify dominant narrative frames, revealing that the majority of top influencers represented healthy bodies as vessels of cultural consumption. She first-authored a conference paper on the results and credits this research as central to her acceptance into the Health Informatics PhD program at UIUC. She is now a tenure-track professor at SUNY Korea in the Department of Technology, AI, & Society.

Kaycee Stiemke was a cognitive science student at UCLA interested in productive and performative activism on social media in the Black Lives Matter movement. I trained her to mine 300,000 Instagram posts surrounding the George Floyd protests and to use Python for data visualization and time-series modeling. Her analyses revealed how performative activism from white influencers drowned out grassroots organizing during the protests. She presented this work at conferences across the United States and is building on these studies as a PhD student in Human-Computer Interaction at Carnegie Mellon, funded by a GEM Fellowship.

Kedar Gupta was a UCLA psychology student whom I trained in computational cognitive modeling of belief systems. He developed multilevel Bayesian models and fit them to surveys I conducted to predict subject-level effects of different types of educational interventions on death penalty attitudes. This research experience gave him the training needed to join a computational neuroscience lab at Columbia as a fully funded PhD student.

Caleb Musfeldt was an ASU math major who wanted to understand how AI models may emulate human reasoning. Through one-on-one mentoring, I guided him in reading and evaluating research papers and advised him on his graduate school application. I also connected him with colleagues with whom he developed independent research relationships that led to publications. He cites my mentorship as critical to his acceptance as a PhD student in computer science and neuroscience at Princeton.

Courses I Can Teach

My training spans narrative theory and communication, cognitive psychology, mathematics, and computational cognitive science, so I can teach the full quantitative sequence, from introductory statistics to Bayesian multilevel modeling, network science, and NLP. I can also teach cognitive, social, and group psychology, as well as communication and media theory. In my courses, I will integrate theoretical and quantitative topics so students learn how each informs the other in current research. I am also ready to develop new courses that bridge these areas: Data Science for the Behavioral Sciences, where students learn NLP and network analysis by modeling online beliefs and narratives; Mathematical Models of Social Behavior, where students build agent-based and statistical physics models of consensus and test them against network experiment data; and a seminar on Technology and Human Narratives, where students pair media and critical theory with empirical research, then build interactive tools and run small experiments. In every course I design, students build a tool or run a study of their own and document the process in their research portfolios.

Public engagement and service

Outside the classroom, I have given a Lectures on Tap talk in Los Angeles, and applied narrative technology at the Miami (Arizona) Loco Arts festival. I am developing public events in the Phoenix metro that invite residents to share their narratives about emerging technologies and explore how their views sit among those of their neighbors. I have also presented my research to AI industry leaders and startups, and serve as an active reviewer for the Cognitive Science Society conference (2018–2026) and as a journal referee for PNAS Nexus, Cognitive Science, Journal of Experimental Psychology: General, PLOS ONE, and Policy & Internet.