Artificial Intelligence and the Future of K-12 Schools: Concerns About the Uses of AI in Education (Part 1)

I view Computers and AI as a type of gift that humanity has given to itself. This reminded me of the quotations given above. (Did you catch my pun in attributing the second quote about naysayers?)

This and the next newsletter explore some of the current and rapidly increasing uses of AI in our schools. Although I am a strong proponent of using computers and AI in education, I have some concerns that I want to share with you.

In talking about possible futures, I like to start with the past. Life on earth has existed for billions of years (Marshall, 7/14/2009, link):

This is our current “best guess” for the beginning of life on Earth. It is distinctly possible that this date will change as more evidence comes to light. The first life may have developed in undersea alkaline vents, and was probably based on RNA rather than DNA. [The article by Marshall includes a photograph of a large tortoise with a statement, “Galapagos tortoises are the product of over 3 billion years of evolution.”]

More recent publications suggest a still much earlier beginning of life on earth (Wikipedia, 2021, link). During these billions of years, some life forms prospered and survived for very long periods of time, while others went extinct. Homo Sapiens have only been on earth for perhaps 200,000 years or so. During this time, we have developed many new technologies together with the lifestyles that make routine use of this technology. We have populated the earth, and our activities have significantly changed the world. In the past few hundred years, our technology and life styles have led to our current global warming crisis, and also contributed to many species moving toward extinction.

We can examine the various technologies we routinely use, and think about how they are changing us Homo Sapiens as well as other life forms on earth. We might want to think of each technological development as a gift we accept, and not consider what effects it is having on us and on all other life on earth.

Or, we can question what we are doing, and consider more carefully how we educate ourselves (and especially our children) to make wise use of each new technology. We can set goals for ourselves such as improving the quality of life of Homo Sapiens while at the same time preserving nature with its many life forms and beauty.

Most people tend to think about AI as something quite new. They think in terms of the current and still increasing capabilities of computers to solve problems and accomplish tasks that require human intelligence. For a long time, I have thought of this definition as too limited.

Abhijit Naskar presents a much broader view of AI. He was born in Calcutta, India, and now is one of the world’s most famous neuroscientists and an untiring advocate of global harmony and peace. He became a beloved best-selling author all over the world with his very first book, The Art of Neuroscience in Everything. In a recent TEDx talk, Naskar states, “Every machine has artificial intelligence” (Naskar, 5/20/2019, link). He argues that the more advanced a machine becomes, the more advanced its artificial intelligence becomes as well. In his writings and talks, Naskar emphasizes that a machine cannot feel what it is doing and this is a fundamental difference between human intelligence and machine intelligence.

I agree with what he is saying about AI, but I expand on it. For example, I think of the human development of written language as a tremendous step forward in AI. This was followed almost immediately by the creation of schools to teach people (including children) reading, writing, and arithmetic. It was recognized how such knowledge and skills empowered its users—and also empowered the people who owned or could hire readers and writers to work for them. Such education was limited initially to a small number of people. Wealthy people paid for the education of the people they selected. Cuneiform clay tablets from these early years of schools indicate that it took up to twelve years for students to achieve the competency levels deemed desirable in these three subject areas.

Over the past 5,500 years, we have decided that a multi-year opportunity to attend schools teaching reading, writing, and arithmetic is an inalienable right of children throughout the world, and a high percentage of today’s children now have this opportunity.

As an aside, the development of the mind tool called writing and the development of schools to teach children the 3Rs were not based on careful long-term research that studied whether or not this was a good thing for the future quality of life of the students or for humanity.

Here is another example. While you were in school, you learned to do paper-and-pencil long division of decimal numbers. Well, at one time number systems had not yet been invented, and even after these systems were developed it took people about 5,000 years to reach our currently used paper-and-pencil algorithms for long division (Pat’s blog, 11/17/2019, link). I think of long division as a human-developed capability that might well be considered a type of AI.

Over thousands of years, some people began to encounter arithmetic problems that required a great deal of calculation. Researchers worked on developing machines that could aid in this task, with the resulting wide-scale availability and use of mechanical calculators. This proved to be a difficult challenge.

The first tool created specifically for use in mathematical computations was the abacus, likely invented in Sumeria around 2500 B.C. The abacus was a table of successive columns with beads or stones representing a single unit, which could be used for addition or subtraction.
…
In 1642, the first true “calculator” was invented: one that performed calculations through a clockwork-type of mechanism. The Pascal calculator, invented by French inventor and mathematician Blaise Pascal, was lauded for attempting arithmetic calculations previously thought impossible (SchoolMart, 3/29/2017, link).

Here is another important example of a machine-based (but not a computer) type of AI. A number of years ago I spent quite a bit of time visiting a third-grade classroom. One day I was talking with two boys in the class, and I asked them if they could tell me the time. They each looked at the analog clock on the wall, and told me the correct time. A clock is one of the great inventions of humanity, an important example of artificial intelligence that was developed and routinely used long before we had computers.

Next, I asked the boys to tell me how long it would be until morning recess, or lunch, or until school let out. They thought, but could not provide me with answers. Aha! They could tell time, but they still lacked an understanding of time, or time sense. They had learned to use a clock, but their education in both the use and the understanding of their time-knowledge was quite low. As a responsible adult, your understanding and use of time are essential to functioning well in our society. A clock has a very low and limited amount of AI. But, you can use a clock, watch, or other timepiece to make a team that working together is able to accomplish many things that neither of you can do alone.

We want children to learn to understand and deal with the types of problems and tasks adults encounter that require not only knowledge of how to tell time, but also to be able to deal effectively with the time-related problems and tasks that are a routine part of adult life. This observation applies to other technologies, such as AI-using computers.

In the remainder of this newsletter and in the following one, I discuss some education-changing aspects of the many different types of AI being developed for use by electronic digital computers. I raise some questions that apply to all such uses of AI in our school.

To begin, let us consider a very simple example. The least expensive electronic calculators have a type of AI that people can learn to use, one that provides them with arithmetic calculation capabilities that often far exceed those of people using paper-and-pencil algorithms. Moreover, it takes very little time to learn to use such a calculator.

These calculators provide an excellent example of a specific non-human type of AI. One underlying goal of math education is for students to develop number sense and math sense. We know how to help children learn to count and to do simple arithmetic. We know how to teach them to do paper-and-pencil addition, subtraction, multiplication, and division. We can look at adult uses and understanding of math, and explore the level of number sense and math sense they routinely demonstrate.

Math education in early grades includes a major focus on learning about numbers and on learning paper-and-pencil algorithms to do arithmetic. So, here are two basic questions:

1. To what extent does having students develop speed, skill, and accuracy in doing paper-and-pencil arithmetic contribute to their development of number sense and math sense?
2. If we partially substitute electronic calculators for paper-and-pencil arithmetic skills, to what extent does this result in students developing less number sense and math sense?

These two questions are illustrative of the questions that we educators face as we deal with the steadily increasing capabilities of AI in our schools. Many educators and others fear that allowing students to make use of calculators will significantly decrease their level of number sense and math sense. They often give examples of today’s high school graduates who lack number sense in dealing with money and in other situations requiring number sense.

But, such testimonials do not constitute proof. I, personally, can examine a widely accepted set of goals of education, and examine how AI-using computers might affect each goal. My analysis draws on my long career in the field of computers in education. Based on this analysis, I can provide you with my own personal thoughts about some of the possibly good and bad effects of such uses.

I read a lot, and from time to time I encounter research that is relevant to these two questions. But mainly I find articles describing individual teachers who have experimented with various uses of the technology, and then telling stories about the results they observed. I enjoy reading such stories. However, they do not constitute the type of research needed to justify major changes in our schools.

This lack of long-term research to support changes in schooling has been going on for thousands of years. As I look back at some of these changes, I am absolutely convinced that they have made important contributions to the improvement of our schools. This indicates to me that we should continue to make changes based on our best judgement, even though we lack the long-term research to justify the changes. However, we should proceed with caution and gather research evidence as we proceed.

I think it is fun to think about some of the technological progress that has made significant contributions to what goes on in our schools. For example, we initially did not have paper and pens for writing. Hmmm. What research do we have to show that moving from writing on clay or wax tablets to the use of paper and pens actually improved student learning, retention, and use of what they were learning in school? This same type of question can be asked about the myriad of other technology-based aids that have come (and, in some cases, gone) that are used as aids in teaching and learning.

Calculators and arithmetic are just the tip of the iceberg. In each subject area that we teach in our schools, we want students to learn to think and to solve the problems and accomplish the tasks they are encountering or are likely to encounter. Thus, a modern education includes learning to make effective use of the AI-based computer tools that now have become routinely used tools in the various disciplines that students are studying.

The usefulness of a specific AI-based computer tool can vary widely from subject to subject, and the use of some of these tools cuts across many disciplines. The Web, Internet, and search engines provide an excellent example of ubiquitous AI-based computer tools whose usefulness spans all disciplines. Unfortunately, many students still lack the necessary connectivity, computer resources, and instruction needed to integrate the effective use of these tools into the various subjects they study.

You might think that the “powers that be” in a country such as the U.S. would understand this and therefore would have taken the steps necessary to provide all students with effective access to these AI-based computer tools. Prior to the Covid-19 pandemic, the U.S certainly had taken some steps to address this issue. However, the amount of progress occurring over the years has proved to be quite inadequate. Now, the Federal Government and a number of state governments are beginning to move on the task of providing every student with good Internet connectivity and suitable computers for use both in their homes and in their schools.

I believe this to be an excellent first step. But, what about the needed training of the children’s parents, guardians, caregivers, and teachers? What about the development of curriculum and instructional materials that make effective use of such facilities? What about ongoing research designed to determine the effectiveness and the flaws in this overall major change in education?

Here is just one example of the challenge we face. For years, educators have talked about authentic assessment. Now, we need to be talking about authentic assessment in an education where students are making extensive use of connectivity and Web-based materials.

Authentic assessment is the idea of using creative learning experiences to test students’ skills and knowledge in realistic situations. Authentic assessment measures students’ success in a way that’s relevant to the skills required of them once they’ve finished your course or degree program (Shaw, 3/26/2019, link).

It now has become routine for people to make use of the Web as an aid in many different aspects of their everyday lives. Thus, as students make more and more use of AI-using computers in their schooling, we certainly must assess this use in the coursework they are doing. Doing so would be a huge change in the way we currently assess students that now is based primarily on the use of closed book, closed notes tests. A small step away from this was a change in the SAT (Scholastic Aptitude Test) that often is used as an aid to determining college admission. The SAT math test has allowed the use of certain types of electronic calculators since 1994. Over the years, the level of capabilities of allowable calculators has been increased, although the currently allowed calculators are far from state of the art.

Humanity has long made use of aids to our physical capabilities. As individuals, we have all grown up making use of machines whose physical capabilities exceed our own in many different endeavors.

We now have AI-using computers whose capabilities exceed our own cognitive capabilities in solving a wide range of problems and accomplishing a wide range of tasks. The capabilities of these AI-using computer systems are increasing steadily. Whether such computer systems eventually will have capabilities that exceed human intelligence across areas in which we make use of our intelligence is a fun question to debate, but is not particularly relevant to the discussion in this and the next newsletter.

Our school systems need to deal with the situation of computers becoming increasingly more capable than are people in many of the problem-solving tasks that humans can use their human intelligence to solve or attempt to solve. The number and importance of these problems will continue to increase.

Thus, our schools need to be providing students with an education for adult life in a world that is routinely using these computer capabilities.

Marshall, M. (7/14/2009). Timeline: The evolution of life. New Scientist. Retrieved 3/5/2021 from https://www.newscientist.com/article/dn17453-timeline-the-evolution-of-life/.

Naskar, A. (5/20/2019). Neurons giveth, neurons taketh away. TEDx Talks. (Video, 15:19). Retrieved 3/4/2021 from https://naskarism.wordpress.com/2019/05/20/ted-talk-by-abhijit-naskar/.

Pat’s Blog. (11/17/2019). Some notes on division, and its history. Retrieved 3/10/2021 from https://pballew.blogspot.com/2019/11/some-notes-on-division-and-its-history.html.

SchoolMart. (3/29/2017). The calculator: A brief history. Retrieved 3/9/2021 from https://www.schoolmart.com/2017/03/29/the-calculator-a-brief-history/.

Shaw, A. (3/26/2019). Authentic assessment in the online classroom. Wiley Education Services. Retrieved 3/12/2021 from https://ctl.wiley.com/authentic-assessment-in-the-online-classroom/.

Wikipedia (2021). Earliest known life forms. Retrieved 3/12/2021 from https://en.wikipedia.org/wiki/Earliest_known_life_forms.