Article: Why AI Should Never Replace Teachers

Photo Source: Raw Pixel

With fall approaching later in September, most children have already returned to school. Artificial Intelligence (AI), machine learning (ML), and large language models (LLMs) such as chatGPT have inspired many pundits and technologists to trumpet the end of many traditional jobs, including the essential work of teachers educating the next generation of citizens in countries around the globe. Some believe that personalized AI-powered tutors will significantly improve education due to the tutorial approach to learning that LLMs could provide. For example, Professor of Computer Science Stuart Russell from UC Berkeley recently suggested that “that personalised ChatGPT-style tutors have the potential to hugely enrich education and widen global access by delivering personalised tuition to every household with a smartphone.” (theguardian.com) Russell remarked that the tutorial teaching style has proven successful at Oxford and Cambridge Universities in the UK. Still, that type of training does not scale because there are not enough teachers to go around for one-on-one tutoring. He notes that ChatGPT-type tutors accessible through smartphones can reach around the globe, making education available to everyone everywhere. Such a vision suggests that simply access to technology will transform education. However, learning involves more than access to information through a screen. Research on learning following the large-scale transformation from classroom to remote learning during the pandemic suggests technology may not sufficiently replace classroom teacher-student interactions. The US Department of Education found that “Data collected before and during the COVID-19 pandemic have shown that in-person learning, on the whole, leads to better academic outcomes, greater levels of student engagement, higher rates of attendance, and better social and emotional well-being, and ensures access to critical school services and extracurricular activities when compared to remote learning.” (ed.gov) Moreover, schools provide a safe and structured environment for children that may not exist during remote learning if parents need to leave the home for work or other responsibilities. As students progress through elementary, junior high, and high school, they encounter different teachers with different teaching styles and perspectives. Having a chatbot teacher, even if it does modify its behavior to adapt to individual learning styles, will miss the opportunity for children to meet and interact in the classroom with different people from their community. The danger exists that a chatbot may express a narrow viewpoint that will apply across the board, limiting growth opportunities and alternative perspectives. With the emergence of large language models such as ChatGPT, technologists suggest that in the near future, children may have an AI-driven teacher-bot that will provide a personalized tutorial learning experience instead of the traditional teacher leading a class full of children through their lessons. OpenAI, the maker of ChatGPT, already offers a tutorial for teachers to incorporate their model in the classroom. (openai.com) OpenAI describes in Section D how a teacher can make an AI tutor for their students to use. It may not represent a full-time replacement for human teachers, but it does introduce the computer as an authority and focal point of student attention. It may seem more straightforward to offload some of the responsibilities of teaching to an AI-driven tutor. Still, the lessons from the effects of remote learning on student progress and the richness of human interaction between students and teachers should make us step carefully when considering replacing teachers with machines.

Dr. Smith’s career in scientific and information research spans the areas of bioinformatics, artificial intelligence, toxicology, and chemistry. He has published a number of peer-reviewed scientific papers. He has worked over the past seventeen years developing advanced analytics, machine learning, and knowledge management tools to enable research and support high-level decision making. Tim completed his Ph.D. in Toxicology at Cornell University and a Bachelor of Science in chemistry from the University of Washington.

You can buy his book on Amazon in paperback and in kindle format here.