June 6, 2017 – How to prepare the next generation for jobs in the AI economy. Read more.

by David KosbieAndrew W. MooreMark Stehlik, Jun 5, 2017

Most of us regard self-driving cars, voice assistants, and other artificially intelligent technologies as revolutionary. For the next generation, however, these wonders will have always existed. AI for them will be more than a tool; in many cases, AI will be their co-worker and a ubiquitous part of their lives.

If the next generation is to use AI and big data effectively – if they’re to understand their inherent limitations, and build even better platforms and intelligent systems — we need to prepare them now. That will mean some adjustments in elementary education and some major, long-overdue upgrades in computer science instruction at the secondary level.

For example, consider how kids are currently interacting with AI and automated technologies: Right now, it might seem magical to tell Siri, “Show me photos of celebrities in orange dresses,” and see a photo of Taylor Swift pop up on a smartphone less than a second later. But it’s clearly not magic. People design AI systems by carefully decomposing a problem into lots of small problems, and enabling the solutions to the small problems to communicate with each other. In this example, the AI program divides the audio into chunks, sends them into the cloud, analyzes them to determine their probable meaning and translates the result into a set of search queries. Then millions of possible answers to those queries are sorted and ranked. Thanks to the scalability of the cloud, this takes just a few dozen milliseconds.

Analytics are critical to companies’ performance.

This isn’t rocket science. But it requires a lot of components – waveform analysis to interpret the audio, machine learning to teach a machine how to recognize a dress, encryption to protect the information, etc. While many are standard components that are used and re-used in any number of applications, it’s not something a solitary genius cooks up in a garage. People who create this type of technology must be able to build teams, work in teams, and integrate solutions created by other teams. These are the skills that we need to be teaching the next generation.

Also, with AI taking over routine information and manual tasks in the workplace, we need additional emphasis on qualities that differentiate human workers from AI — creativity, adaptability, and interpersonal skills.

At the elementary level, that means that we need to emphasize exercises that encourage problem solving and teach children how to work cooperatively in teams. Happily, there is a lot of interest in inquiry-based or project-based learning at the K-8 level, though it’s hard to know how many districts are pursuing this approach.

Ethics also deserves more attention at every educational level. AI technologies face ethical dilemmas all the time — for example, how to exclude racial, ethnic, and gender prejudices from automated decisions; how a self-driving car balances the lives of its occupants with those of pedestrians, etc. — and we need people and programmers who can make well-thought-out contributions to those decision making processes.

We’re not obsessed about teaching coding at the elementary levels. It’s fine to do so, especially if the kids enjoy it, and languages such as Snap! and Scratch are useful. But coding is something kids can pick up later on in their education. However, the notion that you don’t need to worry at all about learning to program is misguided. With the world becoming increasingly digital, computer science is as vital in the arts and sciences as writing and math are. Whether a person chooses to become a computer scientist or not, coding is something that will help a person do more in whatever field they choose. That’s why we believe a basic computer programming course should be required at the 9th grade level.

Only about 40% of U.S. schools now teach programming and the quality and rigor of these courses varies widely. The number of students taking Advanced Placement exams in computer science is growing dramatically, but the 58,000 students taking the AP Computer Science A (APCS-A) test last year still pales in comparison to the 308,000 who took the AP Calculus AB test. A third of our states don’t even count computer science course credits toward graduation requirements.

The U.S. is woefully behind many of our peer nations. Israel notably has integrated computer science into its pre-college curriculum. The UK has made good progress lately with its Computing at School program and Germany and Russia have leapt ahead as well. President Obama’s Computer Science for All initiative, announced in his 2016 State of the Union, was a belated step in the right direction, but may flounder amid budget cuts proposed by the Trump administration.

Expanding computer science at the high school level not only benefits the students, but could help the field of computer science by encouraging more students — and a more diverse group of students — to consider computer science as a career. Though we were thrilled last fall when almost half of our incoming first-year class at Carnegie Mellon was female, the field of computer science is still struggling to increase the number of women and minorities. Engineering intelligence into systems, and finding insights in a ubiquitous sea of data, is a task that cries out for a diverse workforce.

To be successful, however, it is critical that we update the way programming is taught. We’re too often teaching programming as if it were still the 90s, when the details of coding (think Visual Basic) were considered the heart of computer science. If you can slog through programming language details, you might learn something, but it’s still a slog — and it shouldn’t be. Coding is a creative activity, so developing a programming course that is fun and exciting is eminently doable. In New York City, for instance, The Girl Scouts have a program that teaches girls to use Javascript to create and enhance videos — an activity that kids already want to do because it’s fun and relevant to their lives. Why can’t our schools follow suit?

Beyond 9th grade, we believe schools should provide electives such as robotics, computational math, and computational art to nurture students who have the interest and the talent to become computer scientists, or who will need computers to enhance their work in other fields. Few U.S. high schools now go beyond the core training necessary to prepare for the APCS-A exam, though we have a few stunning success stories — Stuyvesant High School in New York City, Thomas Jefferson High School for Science and Technology in Alexandria, Virginia, and TAG (The School for the Talented and Gifted) in Dallas, among others. These schools all boast committed faculty members who have a background or training in computer science.

We also urge high school math departments to place less emphasis on continuous math, including advanced calculus, and more on the math that is directly relevant to computer science, such as statistics, probability, graph theory and logic. Those will be the most useful skills for tomorrow’s data-driven workforce.

A major hurdle is that our schools face a severe shortage of teachers who are trained in computer science. This is where U.S. tech companies could help immensely. Microsoft, for instance, sponsors the TEALS program, which pairs computer professionals with high school teachers for a few hours a week. But we need thousands of educators teaching millions of students. Even greater commitments will be necessary going forward. On the academic side, The University of Texas at Austin’s UTeach program is a model for preparing STEM teachers and has expanded to 44 universities in 21 states and the District of Columbia.

Much more is needed. As with science and math, we need governmental standards driving K-12 computer science education, along with textbooks, courses and ultimately a highly trained national cadre of computer science teachers that are tied to those standards. The Computer Science Teachers Association has been a leader in this area, promulgating a standards framework and an interim set of standards.

Investing in how the next generation understand and interacts with big data and AI is an investment that will pay off in the long run for all of us.