One of the ideas around human-computer interactions that resonated heavily with me in this Gen AI world was the idea of human plus AI is greater than human or AI alone. I remember this was a core idea presented by Sean Nufer during his keynote address at the OU Online Conference on Oct 2, 2024. Elegant and effective. Certainly memorable.
I’ve been thinking more and more about the role of AI in the field of instructional design and educational technology, and one extension of the human+AI idea that I’m wrestling with is Optimal Human Input. Keeping a human-in-the-loop is critical since it leads to better overall AI outputs. In practice, this can include providing detailed prompting, an example of a finished product to model outputs, or repeated rounds of feedback. Each of these takes time and thought, often more than completing the task ourselves. Balancing the time and thought required with the capabilities of generative AI systems is where the idea of Optimal Human Input sits for me. What does this balance look like for educators? What does it look like for the course design process? How do we maintain quality of courses produced in collaboration with AI tools? And how do we truly take advantage of the human+AI paradigms in our field? In other words, to me, Optimal Human Input is the ideal point where humans+AI achieves its greatest outcomes. Applied to educational practice, I find this to be a particularly interesting, outstanding, existential question.
As a quick aside, I’m not the first to use the term or consider similar ideas. Optimal Human Input has been written about by Johannes Schneider in his 2020 paper: Humans learn too: Better Human-AI Interaction using Optimized Human Inputs. That being said, I arrived at these thoughts through my own lens and expertise in the learning design space where my focus is different.
Thinking about Optimal Human Input, let’s consider course design for a moment. What does this look like for subject matter experts? One of their greatest strengths is being embedded in their field. Consequently, a subject matter expert should be engaged in curating and approving the resources to be used in instruction. This is an optimal use of their time and energy to help ground the course materials and provide a strong foundation. Similarly, subject matter experts should specify the types of assessments that promote critical thinking in alignment with the field—including exercises that are representative of industry practices. These are just a couple examples of the value brought by subject matter experts to the course development process and are critical for establishing the direction of the design. Bringing in AI agents to build off this foundation would arrive at stronger outputs than instructing AI tools to derive these components themselves.
Subject matter experts are just one contributor of the instructional design process, and applying Optimal Human Input to their role seeks to ensure the human-in-the-loop is achieved in ideal conditions. We could delve into each contributor to course development and the role they should play in accordance with Optimal Human Input. Briefly, an instructional designer might prepare the aligned learning outcomes for the course in accordance with the program learning outcomes, while an educational technologist may provide insights for supporting specific tools like those for facilitating student podcasting or blogging. Similarly, a media expert may provide the filming opportunities and support to instructors creating their own video materials, while a quality assurance expert might provide the review of human and AI materials alike. All roles are critical to this process, and it is our duty to explore and determine what Optimal Human Input looks like in our individual environments to provide fulfilling learning experiences to our students.
Generative AI can perform miraculous tasks, such as copying course components into learning management systems, producing practice exams, checking outcome alignment, proposing real world applications of course content, prototyping entire modules, editing scripts, editing videos, developing assistive software, scheduling tasks, assisting with project management; the list could go on and on and on. The point is that we have entered an age of collaboration with generative AI systems. Consequently, we should be intentional about our usage of AI tools, identifying where we want to maintain and optimize our human input to provide the best outputs possible for our students.
In accordance with human+AI I invite you to critically consider your own interactions with AI systems and what work you would be willing to offload or might make sense to augment. I consider this to be an existential question about my purpose and what matters to me in my work. Understanding our own Optimal Human Input will help us make deliberate decisions of the role generative AI will play in our professions. And maybe, just maybe, this will help us hold on to our humanity in the long term.