I have been exploring and writing about AI now since the release of ChatGPT. I think about AI both as a user and as an educator and believe I have enough experience and a perspective from which to offer an overview. The tools have improved to a level at which I find them quite valuable as a research and writing assistant. My main approach is to use a tool that allows me to interact with my own digital notes collected over maybe a decade to assist in many of my writing projects. I count myself among those who advocate for the professional value of AI.
My thoughts on AI in education are more complex and it is this perspective I will try to share here. Let me start with what I think is an essential assumption and that is students will find ways to use AI. Whatever the perspective of individual educators is, I argue it is necessary to begin accepting this assumption. Trying to think across many different content areas and skills, it seems reasonable to stipulate that there are certain skills that must be practiced directly to develop (e.g., writing, reasoning, problem solving) and elements of information that may not be life changing whether retained in an individual’s memory or not, but that the general benefit of existing knowledge which is about stored information and the connections that exist within this information offer advantages in understanding and reasoning. Any given fact can certainly now be searched when needed, but this option does not account for the general benefits of what I would describe as general knowledge. We accumulate general knowledge by interacting with our world, but the purposeful accumulation of important information is more efficient through the process we commonly call education. Let me add one more assumption to this position statement. We cannot learn for others nor can we make them learn. We can at best provide access to information and provide external tasks that have the potential to influence the processes of learning. Ultimate responsibility must be placed on individual learners and this is often a requirement at a time when individual learners lack the background and perhaps cannot make decisions understanding how learning works and how skills and knowledge may influence their futures.
Here is my thinking about AI. Educators have a limited amount of time during which they can directly influence learners. They must depend on the cooperation of learners and perhaps their parents when attempting to influence learning during other times. I would describe this reality as important in making decisions about how this time of maximum influence will be spent. For example, I write a lot about study behavior. Educators and sometimes do use class time for studying. When they make this commitment, they are also reducing the time available for other experiences – presentation of information, experiences such as science labs, peer interactions such as guided discussions and debate. Some reactions to AI suggest that class time be used to some extent to control the use of AI. For example, writing a theme during class rather than at home or study hall or completing math problems during class rather than as homework. If the limitation of AI is determined to be significant enough, this can be done, but this will then replace other activities.
So, I believe that the development of some skills and a general knowledge base cannot be eliminated because of AI and this development can only be guaranteed during the time during which an entire class would have to be prevented from using AI. To be clear, I am not advocating for this option. I am trying to identify the benefits and costs of options which I believe cannot be individualized; e.g., educators cannot really differentiate what is required of different individuals in a classroom situation.
Much of what we are playing with involves decisions about when to attempt to exert control over personal goals and motivations. I was a university prof and there is a common approach at this level that differentiates the requirements for a major from general education requirements. We don’t allow students to decide if they want to develop basic writing skills because we require a couple of semesters of composition. We expect a basic level of function in mathematics, but allow individuals to make decisions as to whether the basic course will be what amounts to a high school level course in algebra or the Introduction to The Calculus.
How strongly do we as educators believe we should ignore personal goals and motivation? This is a question for us and for other stakeholders in the educational process. We certainly cannot control learners, but we can arrange evaluation processes to recognize when some mandatory proficiency has not been achieved. Politicians and the general public already tend to blame educators when basic proficiencies do not match those existing in other countries or when graduates seem unprepared for vocations or for civic responsibilities. What consequences do those who are critical suggest for educators or what are they willing to tolerate for the learners who are ultimately responsible?
When I write about this topic it becomes clear to me that the issues I address are very complex and perhaps that is a useful message for others who have simplistic positions on the process of education or the issues educators face. I am a big fan of research informing practice. One challenge with the type of issues described here is that most involve cumulative effects over an extended period of time. Longitudinal studies may eventually provide useful insights, but the downsides could impact an entire generation before the research makes this outcome clear.
Is a summary possible?
I am willing to say that AI offers great benefits to supplement human actions. We all should be prepared to take advantage and guided experience in developing AI-related knowledge and skills should now be a component of what we teach.
Reliance on AI in place of tasks that develop skills is detrimental. You cannot learn to write if AI replaces your attempts to write. You cannot develop critical thinking or reasoning skills if you do not struggle with tasks that require these skills. The issue then is whether the skills are important to the individual and when is the optimal time to make this decision. Perhaps even this is too narrow of a perspective. What are the commitments each of us owes to each other when it comes to basic knowledge and skills?
If forced to take a position, I would suggest that individuals be required to learn and be able to develop knowledge and skills AI-unaided and be able to demonstrate they can apply AI in ways appropriate to the tasks they presently must accomplish. The notion of tool or augmentation seems useful here and it would seem curriculum developers could differentiate cognitive skills from tool proficiency accordingly.
Note: I find that as I write about this topic I encounter the complexities that I think are important to consider. I certainly welcome comments that address these complexities and possibly provide me when ideas I can address in response.
