Introduction
AI is here to stay. The promise of streamlined workflows, a shorter workweek, and overall convenience in our daily lives has solidified AI’s place in our world. In education, we strive to backwards design curriculum and the codification of necessary skills from the future world our students will find themselves in. Preparing our students for their future is the critical focus of education and has always been our North Star.
Assessment
The growth of AI tools is both challenging the processes by which we engage our students and providing opportunities to effectively and efficiently scale well-established instructional practice (pedagogy). Instructional assessment processes, that is, how we measure knowledge and skill acquisition, have always been a central component of education. Traditionally, educators have worked to balance summative assessment and formative assessment.
The goal of summative assessment is to evaluate student learning at the end of instruction, often leveraging pre-existing benchmarks. This is an end-product, rather than process focused, area of assessment. What does this matter for AI? One example that illustrates AI’s relevance is if the summative assessment in a class is a take-home essay, can an educator truly trust who, or what, is doing the work? When responses were copy/pasted off of a website educators learned how to flag the work but AI tools are increasingly unique responses that are challenging to identify for authenticity.
The exciting contribution AI brings to assessment is in the formative assessment realm. Here the goal is to monitor ongoing student learning and to provide meaningful feedback. It is difficult for an individual educator to give this type of ongoing feedback consistently for large groups of students. AI can be a powerful tool to personalize learning for students with rapid, in-the-moment feedback to nurture their comprehension during their learning process.
Instruction
AI, like all technology, is an amplifier, in an educational setting, AI amplifies both good practice and strategies that need improvement. The Apprenticeship Model, at it’s core, is arguably the most natural, effective form of teaching and learning, and has remained amongst modern humankind’s greatest scalability challenges. We have filled auditoriums with eager learners, designed self-paced courses, and explored countless feedback systems, all to replicate the painfully obvious: a human touch in learning is incredibly powerful but horribly unscalable. Here an expert works individually with a learner to model, guide, provide feedback, challenge, and validate the learners progress.
Although apprenticeship can differ widely from one context to another, they typically have some or all of the following features1:
- MODELING – the educator carries out the task, simultaneously thinking aloud about the process, while the learner observes and listens.
- COACHING – as the learner performs the task, the educator gives frequent suggestions, hints, and feedback.
- SCAFFOLDING – the educator provides various forms of support for the learner, perhaps by simplifying the task, breaking it into smaller and more manageable components, or providing less complicated equipment.
- ARTICULATION – the learner explains what they are doing and why, allowing the educator to examine the learner’s knowledge, reasoning, and problem-solving strategies.
- REFLECTION – the educator asks the learner to compare their performance with that of experts, or perhaps with an ideal model of how the task should be done.
- INCREASING COMPLEXITY AND DIVERSITY OF TASKS – as the learner gains greater proficiency, the educator presents more complex, challenging, and varied tasks to complete.
- EXPLORATION – the educator encourages the learner to frame questions and problems on their own, and in doing so to expand and refine acquired skills.
While this tried and true method of learning is effective in a 1:1 scenario, it has historically proven a challenge to scale to a larger student audience. AI has the potential to make the apprenticeship model a highly effective pedagogical practice. The ai-apprenticeship model finds the educator MODELING a task for their students (while explaining their thought process) then directing their students to a self-guided, AI-powered, series of tasks which are responsively-enabled to appropriately challenge the learner to achieve expertise (COACHING, SCAFFOLDING, ARTICULATION, REFLECTION, and INCREASING COMPLEXITY AND DIVERSITY OF TASKS). The educator has been enabled to work 1:1 with students requiring support beyond what the AI is offering. Finally, with developments in AI students now have a new set of tools that they can use to produce AI generative products, via question prompts, that represent their content-area expertise (EXPLORATION).The entire time an authentic discourse is imperative to humanize the process with front of classroom check-ins and student led summary dialogue.
Further, we are hearing of innovative educators modeling AI applications through transparent examples with their students, identifying the pros and cons of the technology as it exists today. The modeling approach, where the educator speaks openly about how they use AI, and even explain their reservations, is a life-skill lesson on having grit, or agency, to persevere in overcoming a new challenge. By demonstrating a level of stick-to-it-tiveness that we all need in life, educators are using AI as a use case that has provided opportunities for their students to have clear access to the skill of agency in action.
Safety and Security
AI is not a replacement for teachers. Teachers are essential for providing students with the guidance and support that they need to succeed. AI can be a valuable tool for teachers, helping them scale good practice and effectively reach more students. But we need to proceed with care when deploying emergent technologies into schools, here we are appropriately hyper-sensitive on how these technologies will impact our students. Regarding AI, the pedagogy will get worked out, teachers are masters of change, that’s not to undermine the significance of the instructional shift, but it can be done. The challenge we need to resolve is centered on safety and security. Data privacy is a fundamental concern that has to be resolved and better understood. Learners need to ask questions, work through problems, and create less than perfect solutions, all of which are not meant for the public stage, their privacy matters and must be protected.
Change is uncomfortable and sometimes scary but change whether forced (ex. COVID teaching), or by choice, has the potential to guide us to make the world a better place for our children.
Also published: Shippee, Micah, AI in Education: a focus on pedagogy (November 30, 2023). SSRN: https://ssrn.com/abstract=4649752 or http://dx.doi.org/10.2139/ssrn.4649752
Sources
- Collins, A. (2006). Cognitive apprenticeship. In R.K. Sawyer (ed.), The Cambridge handbook of the learning sciences (pp. 47-60). Cambridge, England: Cambridge University Press.
- Collins, A., Brown, J.S., & Newman, S.E. (1989). Cognitive apprenticeship: Teaching the crafts of reading, writing, and mathematics. In L.B. Resnick (Ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser (pp. 453-494). Hillsdale, NJ: Erlbaum
- Ormrod, J. (2019). Human learning (8th ed.). Upper Saddle River, NJ: Pearson.
