Automation and Innovation: A Human-centered Approach to AI for Feedback, Assessment and Learning
Dr. Scott Tonidandel
Avoiding Over-reliance on AI, Human-Centered Approaches, Critical thinking, careful oversight, AI assessment, AI feedback, AI automation, student feedback, meaningful feedback
Dr. Scott Tonidandel’s journey with artificial intelligence is a tale of deep engagement, creative experimentation, and continuous evolution in both research and education. Over the years, Dr. Tonidandel has explored AI from multiple angles, blending traditional deep learning techniques with the latest advances in large language models to address real-world challenges in human assessment and learning.
Dr. Tonidandel’s initial foray into AI was driven by a desire to streamline labor-intensive processes. In his research, he focused on the automated scoring of interviews and assessment centers—tasks that once required significant human intervention. By harnessing early deep learning models such as Birch, moving through Roberta and LSTM models, and eventually transitioning to modern large language models, he transformed the way candidate responses are evaluated. His approach aimed to strip away extraneous information, like the identity of the company a candidate previously worked for, to ensure a more objective and fair assessment. In doing so, he sought to reduce biases that often creep into traditional, human-scored interviews. This method not only increased efficiency but also promoted fairness and inclusion by focusing solely on how candidates handled specific scenarios rather than extraneous details.
In the classroom, Dr. Tonidandel applied his AI expertise to tackle another persistent challenge: providing timely, meaningful feedback to students. Faced with large classes and the difficulty of grading essay responses swiftly, he developed an innovative AI-driven tutoring system. This system was built by training a customizable large language model on the very chapters of the textbook his students were using. Through meticulous prompt engineering, Dr. Tonidandel ensured that the model could generate practice essay questions and offer feedback on the students’ responses. The goal was to give students the opportunity to practice and refine their essay-writing skills without waiting for delayed instructor feedback.
Yet, even with carefully designed instructions, the AI occasionally surprised him. Despite being explicitly told to stick to the textbook’s content, the model sometimes ventured into areas not covered in class. For instance, when a student asked about a specific leadership theory, the AI would provide insights from other texts. This unexpected behavior led to a series of humorous yet thought-provoking exchanges, as Dr. Tonidandel debated with the AI about the boundaries of its knowledge. While such instances highlighted the challenges of ensuring that AI strictly adheres to predefined constraints, they also underscored the evolving nature of these tools and the need for continuous refinement.
Dr. Tonidandel also observed that students were quickly adapting to these AI tools. Rather than relying on traditional online resources provided by textbooks or publishers, many students turned directly to platforms like ChatGPT for generating multiple-choice questions and seeking clarifications. This shift is emblematic of a broader transformation in education where technology not only supplements traditional learning but sometimes even replaces established resources. It challenges educators to rethink their roles—from being the sole providers of knowledge to becoming curators who guide students through an ever-expanding digital information landscape.
Looking ahead, Dr. Tonidandel envisions AI taking on an even more transformative role in both research and education. In research, he anticipates that AI will increasingly automate tasks that are currently manual and labor-intensive. For instance, qualitative literature analysis, which traditionally requires researchers to sift through hundreds of articles, might soon be expedited by algorithms capable of extracting and summarizing key information. In education, while AI tools can provide on-demand tutoring and feedback, Dr. Tonidandel stresses the importance of ensuring that students do not become overly reliant on them. Instead, educators should use AI to enhance critical thinking and to help students understand the underlying principles of the technology they are using.
Dr. Scott Tonidandel’s experience with AI is a reflection of a dynamic and evolving field—one that is reshaping the way research is conducted and how students learn. His work demonstrates that while AI offers powerful tools to automate and innovate, it also requires careful oversight, continuous improvement, and a balanced approach that preserves human insight and critical thinking.