Artificial Intelligence and the Future of K-12 Schools: Part 4: The Future of Education in a World of Artificial Intelligence

As I eat breakfast in the morning, I use my tablet computer to browse the news. Here are a few headlines from my morning’s (2/7/2021) browsing. For me, they provided a snapshot of some major aspects of current and future challenges faced by our educational systems.

Such headlines are designed to catch the attention of ordinary readers who like to know what is going on in their country and in the world. They want us to understand the types of news content being made available to the general public, and to think about the implications of this news for their own future, the future of people they know and care about, humanity, and the world.

I believe all educators need to be aware of the first item on the list. Progress is occurring in developing a brain implant that can be used for a direct connection between a brain and a computer. Wow! That sounds like “way out” science fiction to me. Right now I can speak to my computer. In the future it seems likely that people will think to their computer (Lewis & Stix, 7/17/2019, link).

In this and the next newsletter, I present some of my thoughts about needed educational changes being brought about the increasing use of AI in education, and some forecasts about such changes. My focus is on schooling in the United States. The U.S. and a number of other countries have the wealth and infrastructure necessary to provide all of their students with reasonably good access to computers, the Internet, the Web, and a variety of instructional materials designed to make effective use of such computer resources.

Here is a fundamental idea. Our schools have thoroughly integrated the idea of teaching students to read, and then having students learn to make effective use of reading to learn across the curriculum. A somewhat similar thing is beginning to happen with the use of AI. My forecast is that this will prove to be as big a change agent in education as was the development of reading, and (in my opinion) will become a commonplace, major component of schooling in the future.

Computer futurists use the term technological singularity (or, simply the word singularity) to refer to the time when computers will become more intelligent than humans. Such futurists suggest that this may happen as soon as about ten years from now, some forecast it will occur much later in this century, and still others argue that we are a very long way from this ever happening. However, all agree that in an increasing number of situations computers will develop a type and level of intelligence that can solve problems and accomplish tasks beyond the capabilities of humans (Kurzweil, 11/3/2019, link).

Thus, it behooves all of us to examine each goal of education to better understand how it is and/or could be affected by the level of AI already available or likely to become available in the current decade.

From time to time I try to wrap my mind around what it might be like to live after a technological singularity. I am somewhat calmed by the realization that during my entire lifetime I have lived with machines with physical capabilities that exceed mine. I also have lived with a wide range of tools such as radio, television, audio and video recording, the Internet and the Web, and so on.

Some people fear that AI-using computers will become self-aware, and then decide to take over and run the world. Some fear that such computers will be harsh dictators, or even decide that humans are no longer needed. They sometimes appear to see such computers as potential harsh dictators. So far I have not seen any indications that this will occur, and it is not one of my personal fears. My primary fear today is that people whose values are far different than mine, many of whom have a current focus on usurping powers that I strongly believe belong to the people, will be able to make use of computers and AI to achieve their goals.

TPrehumans and humans have long developed and used tools that aid and extend their physical capabilities. Also, through evolution and their own work, they have developed tools such as oral and written languages, telephones, and many other aids to their physical and cognitive capabilities.

I imagine that the first prehumans who developed rock knapping to make very sharp edges on small stones did not think about the long term consequences of their invention. A similar statement holds for many of the tools that have become a routine part of our lives. So, why should we be concerned about the possible long term effects of learning about and using AI-based tools in schools and in other part of our lives?

It is not new for people to be concerned about what is going on in their children’s schools. Consider the variety of printed materials now made available to students while they are in school. Most schools are very careful not to provide their students with library materials and textbooks that the general population in their community might find offensive.

And, what about tools? The simple handheld calculator provides a good example to consider. Adults certainly accept that it is perfectly fine for adults to make use of calculators in dealing with arithmetic problems they encounter. However, parents and schools continue to believe that learning paper-and-pencil arithmetic is a very important goal of math education in the schools. A modest counter trend is illustrated by the fact that the widely use SAT exam has allowed certain types of calculators to be used in the math part of the exam since 1994. Gradually, as more sophisticated calculators were developed and came into widespread use, the types of calculators allowed on the exam changed. Graphing calculators are included in the current list of allowable calculators (DeGeurin, 8/9/2019, link).

This example illustrates a key idea. We want to provide students with an education designed to help prepare them for everyday life, work, and leisure as responsible and contributing adults. In addition, we know that an appropriate education contributes greatly to an adult’s quality of life and to the overall success of the adult’s country. Thus, we should look at each current goal of education, how it contributes to meeting these educational expectations, and how it is being affected by the ongoing changes computers and AI are bringing to our world and to its education capabilities.

Next, think about the Web that now has become a ubiquitous and everyday tool used by essentially all students in the U.S. and other economically developed countries. I like the analogy between use of a calculator for arithmetic and many math-related tasks, and use of the Web as an interdisciplinary tool useful in all areas of academic inquiry and study. Access to and use of networked computers is now a routine part of the school day for a great many students, and this number is steadily increasing. Since AI is used in much of  today’s software, all computer users are making routine use of AI.

The issue of the use of calculators on exams such as the SAT is actually modest when compared with the issue of the use of the Web when taking school exams. You certainly know about the idea of open book and/or open notes exams. Some teachers make use of such exams. But, what about open networked-computer exams? Do you know any teachers who make routine use of this type of exam?

For me, the issue is one of authentic assessment (Mueller, 2018, link). Authentic assessment has long been an important idea in education. As computers become a routine part of the lives of all students, schools need to develop assessments in which it is expected that students will make use of computers in the assessments.

Authentic assessment includes testing students’ abilities to use the “tools of the trade” in solving the types of problems and accomplishing the types of tasks that we want students to learn about in their education. This is and will continue to be a major challenge in education, because the capabilities of AI will continue to improve significantly year after year. It will be fun to see how we progress in addressing this challenge in the coming years.

The above comments are some of the types of thoughts that go through my head as I consider current and possible future roles of AI in schooling. Here are three basic questions I ask myself:

1. What are some routinely used AI-based tools in our current schools, and are these uses advancing the schools’ goals for education?
2. What are some currently available AI-based tools that might be quite suitable for use by students and teachers, tools that currently are either not being used at all or are used only lightly?
3. What are some AI-based tools currently being developed or envisioned that have the potential to have a significant impact on schools?

The remainder of this newsletter and the following one will explore these questions.

Thinking about the three questions listed above leads me to go back and consider the general purposes and uses of computers. The first electronic digital computers were tools developed to help solve problems and accomplish tasks that humans working with the other tools available at that time could not accomplish. Arithmetic calculations provide an excellent example. The first commercially available computer in 1952 could perform 1,000 arithmetic operations per second. At that time, people had many problems they wanted to solve in which such computational speed was very helpful. As computer speeds and availability increased, the discipline of business data processing developed. Routine learning about computer use in data processing—a use that made only very rudimentary use of AI—became commonplace and so became a subject taught in high schools.

During the nearly 70 years since 1952, the speed of the fastest commercially computers has been increased by a factor of about 400 million million (400 times 1012). Researchers and computer programmers have learned to make use of that speed to accomplish many tasks that are far beyond the capabilities of humans.

Let’s look at just one of these tasks—the storage of information on the Web and its retrieval using a browser. Each of these tasks makes extensive use of AI. It is not necessary to have any knowledge about AI in order to use these capabilities effectively. This is an important idea about AI, and I like to think about various analogies to explain what I mean. For example, a person does not need to know the details of building a car in order to drive one. This holds true for many of the routinely used tools humans have developed.

It boggles my mind to think of how I can use AI to search through billions of online documents in less than a second! Yet this is only one example of the many applications of AI that now are an accepted component of schooling. However, our school systems have made only modest progress in dealing with the possible changes in content, pedagogy, and assessment that I believe need be made based on these computer capabilities.

My forecast for the future is that schools will still be struggling with these challenges for many years to come. I currently rank the effective integration of computers and AI into the K-12 curriculum as the most important challenge that schools have faced since reading and writing were first developed.

I assume that all of my readers have concerns about the future of jobs and employment. In IAE Newsletter #297, the second newsletter in this current series, I discussed a broad list of goals of education. Currently, a unifying goal of education is to prepare students for gainful employment. Such education is viewed by many as a vehicle to getting a well-paying job, and also contributes to building a strong, economically successful country.

This emphasis on preparing students for jobs, when combined with the fact that an increasing number of available jobs now require the skilled use of computers, forces us to pay a great deal more attention to the types of computer knowledge and skills currently being taught in our schools. In making forecasts about the future of AI in education, I assume that a unifying purpose of education for gainful employment will continue to be a major and appropriate unifying goal for schools.

However, intelligent computers and automation may well eventually change this emphasis on preparation for wage-earning jobs. Many science fiction writers envision and explore future worlds in which schooling no longer focuses on preparation to obtain and hold jobs. They forecast a world in which all people receive a level of income from the government that is deemed to be enough to support an adequate standard of living. In such stories, nobody has to work in order to have a livable level of income. The authors typically posit a wide variety of higher paying jobs being available for those who are qualified to do the work and want to have the additional income.

Today in the U.S., we already have very large scale examples of government payments to some segments of our population. We have Social Security payments and Medicare for people of retirement age. We also have various forms of supplemental income payments for people with either no income or one that is totally inadequate, and for people with certain types of disabilities. We continue to explore possible ways to deal with the need for all people to have a good quality of life, one that includes access to adequate medical care. In a number of countries, the costs of medical of care are paid by the government.

The length of the work week and minimum rates of pay are another quality of life issue. Over the past 200 years, the average length of the work week for fulltime employees in the U.S. has decreased from being 72 hours or more to 40 hours (Sawyer, 12/25/1977, link). This decrease in the average length of the work week certainly has helped to improve our quality of life.

Fourteen currently widely accepted goals of education were presented and discussed in IAE Newsletter #297. Preparation for employment, one of the unifying ideas in this list, was introduced in the section above. My focus here is on the first of the 14 goals, with the next newsletter examining how AI will strongly affect some of the other goals.

G1. Information and Communication Technology (ICT): All students have appropriate knowledge and skills for using our rapidly changing ICT as well as related technologies relevant to their lives and our world.

Dick Ricketts and I wrote about this goal more than 30 years ago in our book, Long-range Planning for Computers in Schools. This was by no means a new idea at that time. In the early 1970s some leaders in the field of computers in education proposed that all students needed to become computer literate.

These visionaries already had seen how computers empower their users, and that quite young students could learn to use computers at a level beneficial to them. The visionaries made use of an analogy between students becoming reading and writing literate, and becoming computer literate. For many of the visionaries, computer literacy includes having students learn to write their own computer programs.

For nearly 50 years, education leaders have been thinking about and making a variety of attempts to address the issue of computer literacy for all students. During this time the cost, availability, and capabilities of computers have changed markedly. In some sense, schools have been playing a catch-up game. Unfortunately, it appears likely this will continue to be the case for many years to come.

I tend to think about this challenge in a number of different ways. Perhaps first is the issue of providing all students with good access to computers and related hardware, the Internet, and the Web, both at home and at school. It is equally important to provide the instruction needed to help all students learn to make effective and appropriate use of such computer facilities. For years, I have raised the question: What do we want students to learn about a type of problem deemed important in the curriculum content, a problem that computers can solve independently and/or greatly help in solving?

I find that an analogy with reading and writing is useful. Over thousands of years, we gradually have determined the levels of reading and writing literacy that typical students can achieve through various amounts of schooling. The advent of the mass produced Gutenberg Bible in 1455 led many people to decide that it was desirable that students be able to read at the level needed to access this now widely available Bible. This provided a motivation for many to support teaching reading and writing in schools. As a steadily increasing number of students learned reading and writing, the benefits to the students and to the societies they lived in became apparent. Thus, one measure of the success of education is the worldwide level of literacy. Education that includes learning to read and write is now considered to be a birthright of most children throughout the world.

There is a substantial difference between being able to read one particular book, and being able to communicate effectively in writing. Desired levels of writing skills—including spelling and grammar—are difficult to achieve. We have not yet decided on the extent to which an educational goal should be for students to write well in a word processor environment that includes a spell checker and grammar checker. Interestingly, we now have some computer programs that  are able to write entire documents based on a list of ideas or topics provided by the user. This brings to my mind an analogy of a business executive who has a highly competent secretary, one to whom he can give a simple outline of the points he wants to make, and the secretary has the skills required to write and distribute to the appropriate recipients a very professional letter covering that list of points. We still have a long way to go before AI can accomplish such a task.

There are some significant advantages of teaching students to use computers as a routinely-used aid to their writing. This raises an interesting question. Who needs to be able to read and write cursive, when keyboarding and computer aids are readily available and so routinely used? A number of schools already have dropped cursive handwriting from their required curriculum.

Here is another challenging question. Do we want students to learn to read and write multimedia? Since the development of the BASIC and Logo programming languages, we have known that quite young children can learn to write computer programs. Moreover, the success of young students in learning the programming language Scratch and other modern programming languages designed for children, and also in doing programming-like activities in a number of computer-based games, provides ample evidence that young students are capable of learning to create multimedia documents. Here are two questions to ponder. Should instruction in reading and writing multimedia be required in schools? If so, should the routine use of these skills become an ordinary, everyday part of the curriculum at all levels?

Children growing up in our country today gain a considerable level of computer literacy independently of any formal instruction in this area that they receive in schools. But, there is much that students can be learning in school about the overall topic of computational thinking. In brief summary, computational thinking is the type of thinking used in analyzing and figuring out how to solve problems in which a computer may be a valuable aid to actually solving the problems. Achieving the effective integration of computational thinking throughout the curriculum is a major challenge. The next newsletter will explore various aspects of this challenge.

DeGeurin, M. (8/9/2019). Here’s how the SAT has changed over the past 90 years and where it might be heading. Insider. Retrieved 2/6/2021 from https://www.insider.com/how-the-sat-has-changed-over-the-past-90-years-2019-8#1980s-2000-sat-places-greater-emphasis-on-reading-and-allows-calculators-5.

Kurzweil, R. (11/3/2019). The future of intelligence, artificial and natural. C Innovation Global. (video, 63 minutes.) Retrieved 2/8/2021 from https://www.youtube.com/watch?v=Kd17c5m4kdM&feature=youtu.be.

Lewis, T., & Stix, G. (7/17/2019). Elon Musk’s secretive brain tech company debuts a sophisticated neural implant. Scientific American. Retrieved 2/7/2021 from https://www.scientificamerican.com/article/elon-musks-secretive-brain-tech-company-debuts-a-sophisticated-neural-implant1/.

Mueller, J. (2018). Authentic assessment toolbox. Retrieved 2/6/2021 from http://jfmueller.faculty.noctrl.edu/toolbox/whatisit.htm#traditionalAuthentic.

Sawyer, K. (12/25/1977). 200 Years Ago – The 12-hour day, the 6-day week. The Washington Post. Retrieved 2/6/2021 from https://www.washingtonpost.com/archive/politics/1977/12/25/200-years-ago-the-12-hour-day-the-6-day-week/8a0f3c78-b7a0-4db4-ac33-00649519d1eb/.