Skip to main content

MEITE: Learning to leverage emerging technology

HDFS alumnus and MEITE student Alexander Sumner, alongside his classmates, developed Promptly, an innovative puzzle game that leverages theory to redefine how audiences understand and engage with AI technology.
Image of Alexander Sumner.

The cutting-edge Master of Arts in Educational Innovation, Technology, and Entrepreneurship (MEITE) program prepares students to leverage the latest technologies, educational research, and design thinking to shape the future of education in a wide range of professional settings. Through rigorous coursework and unparalleled internship experiences, MEITE graduates leave Chapel Hill ready to produce, utilize, and analyze high-impact educational innovations for the social good across sectors.

MEITE student Alexander Sumner (’24 B.A.Ed.) takes us inside the program and discusses Promptly — a class project that leverages AI to enhance AI literacy — developed with MEITE students Aiden Hoke, Mengyao Cheng, Noah Krumme, and Sonny Santiago in EDUC 590: “Generative AI for Learning and Innovation,” a course taught by faculty member Ha Nguyen, Ph.D., in fall 2024.

AI has transformed—and continues to transform—how we live, work, and learn. Yet, as AI tools become increasingly powerful, the gap in AI literacy among Americans continues to grow. Following my HDFS scholarship in the School of Education, the MEITE program has given me the confidence to tackle difficult challenges, like the AI literacy gap. I ask myself, how do we empower learners to navigate AI thoughtfully?

In the course “Generative AI for Learning and Innovation,” our task was to propose and develop an innovative product that addressed a key need in education. The challenge was threefold: to design a solution that integrates generative AI tools, apply insights from user testing, and evolve through an iterative design process.  

Enter Promptly, an AI Literacy puzzle game I had an idea for and designed with my MEITE cohort members. Promptly teaches players the essentials of prompt engineering — a skill at the heart of effective human-AI interaction. The game challenges players to reverse-engineer the precise words used to prompt the style, subject, setting, and mood of the AI-generated Daily Challenge image.  

Initially, Promptly’s purpose was for middle school education. Now, we know it can benefit anyone hoping to make the most of Gen-AI image generation. 

Promptly scaffolds the learning process by providing a CLEAR — Concise, Logical, Explicit, Adaptive, Reflective — prompt structure (Lo, 2023), which helps players break down the complexity of prompt engineering into four manageable elements. Instead of crafting open-ended prompts, players guess specific words, simplifying what could otherwise feel overwhelming. Our structured approach enables learners to intuitively grasp the components of an effective prompt while building their confidence and skills in an engaging way. 

Unlike many popular online puzzle games, Promptly harnesses AI’s profound ability to provide instant, adaptive feedback for users’ attempts. Each guess is scored based on its semantic similarity to the correct answer, with feedback carefully engineered to guide players toward a “Proficient” score (0.800) or the ultimate Promptly score (1.000). By using gamification, Promptly not only teaches critical skills for working with AI but also fosters creative problem-solving and engagement. 

Designing for learning 

In designing Promptly, Dr. Nguyen always asked us, “How are you backing Promptly with learning theory?” My team leveraged insights from what we were learning in class and the program. During the fall semester, I also took Morgan Distinguished Professor of Educational Innovation Dr. Keith Sawyer’s intro to Learning Sciences course, in which I learned about Cognitive Load Theory (CLT). CLT emphasizes the importance of managing working memory, which has limited capacity when processing new information. Effective learning design reduces extraneous cognitive load — or unnecessary distractions — while increasing germane cognitive load, the mental effort directed toward learning and understanding. CLT helped us to address two key issues revealed during user testing: (1) confusion about the game’s rules and interface, and (2) limited engagement with the feedback provided. Our “design challenge” was to reduce extraneous cognitive load so players could learn to play the game effectively and use the game’s feedback to gain experience with AI image generation. 

Initially, we relied on a text-heavy “How to Play” pop-up to explain how the game works. User testing revealed that most participants skimmed or ignored the instructions entirely, leaving them unclear on how to actually play. This confusion reflected an overload of working memory, which prevented users from building the necessary understanding of the game. 

To address this, we implemented an “interactive tutorial,” guided by findings from Reisslein et al. (2006), which says novice learners benefit most from step-by-step walkthroughs followed by guided practice. Our revised design featured vibrant visual cues and concise instructions, guiding players through one interactive element at a time. This approach reduces extraneous cognitive load by presenting manageable chunks of information, ensuring that users can quickly and confidently learn how to play. 

Another challenge involved players neglecting feedback provided during gameplay. Initially, we provided feedback within a dropdown menu, which many participants failed to notice or found too cumbersome to engage. Research by Tärning et al. (2020) on feedback neglect told us learners often avoid feedback if it requires too much effort to access or process. 

To solve this, we introduced “hovering feedback,” a feature that displays concise, adaptive feedback directly when users hover over their guesses — making feedback noticeable and immediately actionable. By providing incremental, easy-to-interpret feedback tailored to each attempt, we reduced the extraneous cognitive load. The incorporation of visual cues, such as a “guess meter” and side-by-side image comparisons, further reinforces users’ understanding and progress while keeping them engaged. 

 

By incorporating these evidence-based design solutions, we ensure that Promptly equips learners to navigate and master the process of AI prompt engineering in a way that is both intuitive and engaging. These innovations reflect the power of integrating learning science into game design, making complex concepts accessible to all learners (shoutout Dr. Sawyer)! 

Promptly started off as a mere idea a spark of inspiration in a classroom discussion. Through the fall semester, that spark — with the insights of an awesome team and faculty member — grew into a fully realized educational tool.  

The course taught me to embrace the iterative process: brainstorming, prototyping, conducting user tests, and refining at every step. By applying design thinking principles, I’ve learned how to approach complex educational challenges with creativity and focus. 

Promptly serves as a reminder that, in a world increasingly influenced by automation, we can create educational experiences that preserve human agency and empower learners. When grounded in thoughtful design and research, technology has the potential to serve as a catalyst for growth amplifying creativity, fostering critical thinking, and inspiring innovation. 

+ Share This

By Alexander Sumner