There is one question that almost everyone asks when discussing artificial intelligence: Will it replace human beings, or augment them?
The answer is not as black and white as you may think, especially when you apply AI to one of the most critical, formative careers of all — teaching.
If you ask any teacher why they educate for a living, it is because they want to make a difference. But teaching is becoming an increasingly businesslike profession, with a vast amount of internal and external politics and processes to deal with, such as league tables, standardized testing, and more.
In fact, according to a report from the Learning Policy Institute, 200,000 teachers are leaving the profession each year due to frustrations beyond their control.
Enter AI. HappyNumbers is a personal teaching assistant that uses AI to help K-2 teachers deepen a student’s conceptual understanding of math.
“Teachers want to inspire kids, to change their lives — not to burn out on stress, paperwork, meetings, and other factors that are well-documented,” Evgeny Milyutin, CEO and cofounder at HappyNumbers, told me. “AI in education is an augmentation play because it lets teachers reclaim their learning time and ‘take back’ their profession, which is a calling. Rather than making teaching less human, we believe that AI tools can make it more human again.”
While AI can do a lot of incredible things already, many suggest it is still weak when it comes to emotional intelligence.
“Human relationships are difficult to reproduce,” Milyutin said. “Building a personal and emotional connection is what teachers can do well but technology can’t. This is extremely important for students — especially primary students. However, personally connecting with each student costs time and resources.”
While emotional connections will take some time to replicate, AI is particularly good at one thing: Personalization. As we saw throughout 2017, AI has become almost ubiquitous in personalized marketing technologies, and using those same techniques in teaching offers a great deal of value for teachers.
“Personalized learning is key to encourage higher order thinking,” Milyutin said. “But, at the same time, uncovering what a student is motivated by and pinpointing their interests is also necessary to create critical thinking. Research has shown that knowing what a child values and what they are motivated by may be the key to if they will be able to utilize critical thinking skills in the first place. This is a very personal task that is more suited to human beings than to technology.”
And AI is helping in one particularly important way — illustrating the difference between data and what that data means in the real world. According to a report from the Bill and Melinda Gates Foundation, 86 percent of teachers are continually looking for ways to engage students based on who they are.
“There is a difference between raw data and insights,” Milyutin said. “While students are working, AI can learn a lot of insightful information about each of them that is essential for teachers who aim to provide individual instruction to each student of the class. Teachers can use such insights to drive their actions, making classroom instruction tailored to each student — not in a one-size-fits-all way.”
HappyNumbers has, so far, solved more than 30 million math problems through its platform. While a teacher tutors a small group, HappyNumbers serves as an independent math center, providing individualized instruction for the rest of the class. It is available for iOS, Android, PC, Mac, and Chrome.
The system combines AI, personalization, the latest learning techniques, and psychology to help guide students to success.
“We have seen many apps claiming a fun and engaging experience,” Milyutin said. “The problem here is that it should not be just fun and engaging, but it should be learning. Any video game is fun and engaging, but is there any learning? With HappyNumbers, the students are doing math; it’s not a video game with math as a distraction. We’ve done lots of testing to understand how it should be designed to be interesting to the student.”
Math, famously, is not an easy subject for every student either. Some people can pick it up naturally, but not everyone has a propensity for numbers and logic.
“An important thing to mention here is math anxiety,” Milyutin said. “You should not make the student feel that they are wrong or unable in math. Imagine giving a student 10 math problems; the student solves them, and you report to the student ‘You solved 4 of 10 correctly,’ with no information on how to fix those mistakes. HappyNumbers is exactly the opposite of this. It adapts to the student’s pace and level, and it makes sure that the student is never stuck on the problem, and any mistake the student makes is not perceived by the student as a mistake. Instead, HappyNumbers will provide just enough instruction for the student to understand how to fix such mistake as soon as it’s made.”
So what’s next for HappyNumbers and the use of AI to augment the teaching profession?
“Our team is working on ensuring that students who use HappyNumbers are provided with enough exercises to cover the learning standards to be achieved,” Milyutin said. “Another is to ensure that teachers who use HappyNumbers are provided with actionable insights about their students, so they can better individualize their instruction. Many learning apps/tools report data or information — in order to use it, teachers would need to spend a good amount of time to analyze it and understand what to do with it. Bringing teachers solid, actionable insights instead of raw data/information will save teachers lots of time (that they always don’t have) and will make their job easier and more effective. This is our ultimate goal.”