The Solution to Our Education Crisis Might be AI

Tomorrow's Teachers

Robots will replace teachers by 2027.

That's the bold claim that Anthony Seldon, a British education expert, made at the British Science Festival in September.

Seldon may be the first to set such a specific deadline for the automation of education, but he's not the first to note technology's potential to replace human workers. Whether the "robots" take the form of artificially intelligent (AI) software programs or humanoid machines, research suggests that technology is poised to automate a huge proportion of jobs worldwide, disrupting the global economy and leaving millions unemployed.

But just which jobs are on the chopping block is still a subject of debate.

Some experts have suggested that autonomous systems will replace us in jobs for which humans are unsuited anyway — those that are dull, dirty, and dangerous. That's already happening. Robots clean nuclear disaster sites and work construction jobs. Desk jobs aren't immune to the robot takeover, however — machines are replacing finance experts, outperforming doctors, and competing with advertising masterminds.

The unique demands placed on primary and secondary school teachers make this position different from many other jobs at risk of automation. Students all learn differently, and a good teacher must attempt to deliver lessons in a way that resonates with every child in the classroom. Some students may have behavioral or psychological problems that inhibit or complicate that process. Others may have parents who are too involved, or not involved enough, in their education. Effective teachers must be able to navigate these many hurdles while satisfying often-changing curriculum requirements.

In short, the job demands that teachers have nearly superhuman levels of empathy, grit, and organization. Creating robotic teachers that can meet all these demands might be challenging, but in the end, could these AI-enhanced entities solve our most pervasive and systemic issues in education?

Room For Improvement

In 2015, the United Nations Educational, Scientific, and Cultural Organization (UNESCO) adopted the 2030 Agenda for Sustainable Development, a plan for eliminating poverty through sustainable development. One goal listed on the agenda is to ensure everyone in the world has equal access to a quality education. Specific targets include completely free primary and secondary education, access to updated education facilities, and instruction from qualified teachers.

Some nations will have a tougher time meeting these goals than others. As of 2014, roughly nine percent of primary school-aged children (ages 5 to 11) weren't in school, according to the same UNESCO report. For lower secondary school-aged children (ages 12 to 14), that percentage jumps to 16 percent. More than 70 percent of out-of-school children live in Southern Asia and Sub-Saharan Africa. In the latter region, a majority of the schools aren't equipped with electricity or potable water, and depending on the grade level, between 26 and 56 percent of teachers aren't properly trained.

To meet UNESCO's target of equal access to quality education, the world needs a lot more qualified teachers. The organization reports that we must add 20.1 million primary and secondary school teachers to the workforce, while also finding replacements for the 48.6 million expected to leave in the next 13 years due to retirement, the end of a temporary contract, or the desire to pursue a different profession with better pay or better working conditions.

 

That's...a lot of teachers. So it's easy to see the appeal of using a robotic teacher to fill these gaps. Sure, it takes a lot of time and money to automate an entire profession. But after the initial development costs, administrators wouldn't need to worry about paying digital teachers. This saved money could then be used to pay for the needed updates to education facilities or other costs associated with providing all youth with a free education.

Digital teachers wouldn't need days off and would never be late for work. Administrators could upload any changes to curricula across an entire fleet of AI instructors, and the systems would never make mistakes. If programmed correctly, they also wouldn't show any biases toward students based on gender, race, socio-economic status, personality preference, or other consideration.

But we still have some ways to go before such instructors enter our classrooms.

Education systems are "only as good as the teachers who provide the hands-on schooling," UNESCO claims, and today's robots simply can't match human teachers in the quality of education they provide to students. In fact, they won't be able to for at least the next decade, Rose Luckin, a professor at the University College London Knowledge Lab, a research center focused on how digital media can transform education, told Futurism. Teachers rely heavily on social interaction to support their students and figure out what they need, Luckin continued, and so far no digital system can compete with a human in this realm.

However, it is possible that no robot will ever be good enough to replace teachers completely. "I do not believe that any robot can fulfill the wide range of tasks that a human teacher completes on a daily basis, nor do I believe that any robot will develop the vast repertoire of skills and abilities that a human teacher possesses," Luckin said.

There is some weight to Luckin's assertions. While machines can handle a variety of specific tasks, we haven't yet come close to creating artificial general intelligence (AGI) — the kind of machine that could answer the tough questions outside the purview of the immediate lesson that good teachers should be prepared to tackle. Today's robots also lack the empathy and ability to inspire that teachers bring to the classroom.

That doesn't mean robots won't replace teachers, though. Very few studies directly compare human and robot teachers, so it's not clear how much better the human performs than the robot.

In any case, Luckin suggests a compromise: AI and automated systems could have collaborative roles in the education system. That would enable teachers and students to take advantage of the tech in ways that will benefit them both, and we wouldn't need to worry about lack of oversight for when our AI systems do encounter problems.

AI in Every Classroom

For teachers, the classroom is anything but serene. Kids giggle during lessons, call out, rustle papers, and fidget — teachers must compete with the chaos to simply get students to learn. And teachers take their jobs home with them, too, spending their evenings and weekends planning lessons and grading student work.

What if AI could act as an extra pair of hands in the classroom? That was the idea Luckin and her co-author put forth in a recent paper. This AI assistant could manage tasks such as taking attendance or routine grading. It could also help teachers generate new lessons by autonomously navigating online teaching resources, such as Teachers Pay Teachers, to find the lesson plans most likely to resonate with a classroom based on the details of the students and the school's specific curriculum.

Decreasing the workload dumped on teachers would hopefully make them less stressed. This could limit the burnout that has exacerbated the teacher shortage and make the position more appealing to others considering becoming teachers. In her paper, Luckin predicted that every teacher could have a dedicated AI assistant within the next decade.

 

But AI could do more than the drudgery of teaching — it could actually make teachers better by giving them greater insight into their students' needs.

Classrooms could be equipped with language processors, speech and gesture recognition technology, eye-tracking, and other physiological sensors to collect and analyze information about each student, Luckin writes. Instead of waiting for a test or a raised hand for a student to display her understanding of the material, teachers could access real-time information that could show them why the student might not be learning at full capacity. They'd know which students weren't getting enough sleep, if they had inadequate diets, if they were suffering from emotional stress — information that can affect a student's performance but that can be difficult to tease out in the classroom.

The teacher could use this information to tailor his or her teaching strategies to meet the needs of each individual student. They could simply look at a list generated by the AI to see what each student should work on that day. If a student needed extra one-on-one attention, the teacher could instruct that student to work with an AI-powered Intelligent Tutoring System (ITS) that adjusts its approach to match the student's learning style. Meanwhile, the teacher may decide to give other students group work so they could hone their interpersonal skills.

The system could also help students modify their behavior to improve their own performance. For example, a student might learn that she scores lower on exams when she stayed up late the night before, drank coffee that morning, or took public transport to school instead of walking. By altering these habits in the future, she could score better on the tests and excel.

As Luckin told Futurism, the increased use of AI in education could have downsides. Schools will need to guard against the misuse of student data, and cybersecurity will be of the utmost importance. Still, these types of precautions won't be limited to education. Data protection will be a universal concern as the Internet of Things (IoT) grows and our world gets increasingly "smart."

A few decades in the future, every student and teacher could be the master of their own personal educational analytics, Luckin predicts. That information could be useful beyond the classroom — students may choose to share certain analytics along with their college admission packages, while teachers may include theirs in applications for future employment.

Not My Classroom

As previously noted, it will be a little while before every teacher has an AI assistant and every student has an AI tutor. Developing the necessary technology will be the simple part, Luckin said, and already, some researchers are working on such systems.

Convincing parents, teachers, and students to embrace AI in education will be the real challenge. Some may be biased against the technology for fear it will leave them unemployed, while others may have a hard time shaking thoughts of the doomsday scenarios posited by tech luminaries such as Elon Musk and Stephen Hawking.

Teachers, in particular, are likely to be more resistant to automation because teaching is an "inherently human process," Terry Heick, founder and director of TeachThought, a publisher of teaching materials and resources that are focused on innovation in education, told Futurism.

He said teachers might take the suggestion that a "symbolically mindless robot" could do their work as an indication that others see their skillset as easy to duplicate. In fact, said Heick, the opposite is true — teaching is such an "impossible" task that anything that could make the job easier on teachers should be valued.

https://youtu.be/m6U7JNu8c6Q

To ease this resistance, researchers and tech companies could involve educators, parents, and students in the development of AI systems designed for the classroom, as Luckin suggests in her paper. Stakeholders could see that their input and experience is valuable; they could even identify weaknesses in systems under development. Researchers can tailor systems to suit the needs of educators, students, and parents, improving the final product. Skeptics could see how AI could improve the learning experience.

For this process to go well, it needs to be slow and iterative. It will likely take decades at best.

"Considering that education still hasn’t embraced mobile technology, the idea of Johnny 5 circling around a classroom teaching students in just a decade seems far-fetched," Heick said. But maybe someday.