Education is not rocket science. It’s harder.

GUEST COLUMN | by Elijah Mayfield

Software developers working in education want to help teachers. There’s a lot of software out there that aims to fulfill this goal in the classroom, but the fit isn’t always obvious. At my company, for instance, we make products for writing, and even the simplest essays are a subjective, multifaceted creation.

Can we assign judgment to the diverse decisions of an author? Can any ed-tech really understand the needs of students? Digital tools are, at the end of the day, tools. They cannot replace a teacher’s ability to know the history of a student’s life, their work, and the types of feedback that motivate them to keep trying. Finding a fit in education technology is personal, and that makes software development hard. Most of teaching cannot be automated.

To have the slightest chance of being useful to schools, even the most sophisticated technology product needs to have that voice embedded deep in its design.

This is disappointing to a lot of programmers. In the history of computing, automation has been the easy way to get things done with a computer, and we’ve used them for tasks that have been clean and structured.

• The earliest computers were used in World War II to predict the trajectories of rockets. This is just calculus; the machines were expected to calculate the path of a curve.
• In the early days of artificial intelligence, programmers still gave their programs straightforward tasks. When Kasparov lost to Deep Blue in their chess match, the computer was given a simple job: on a board with sixty-four spaces, pick from one of a limited number of moves. There were clear rules about what is a legal move and what is illegal, and there were no shades of grey.
• This is also true for most business systems in use today. Processing a credit card transaction just requires a card number, an expiration date, a dollar amount. The infrastructure is hard but at least the input is clean.

Software to support writing does not have these luxuries – helping students is a lot harder than getting to checkmate. Student input is messy, their intentions are vague, and the structure of writing is as nuanced as language itself.

But the solution here is not a secret.

Teachers have already discovered countless ways to motivate students. They can make the learning process effective, engaging, and fun. To have the slightest chance of being useful to schools, even the most sophisticated technology product needs to have that voice embedded deep in its design.

So there’s a simple formula that tech companies can follow to get the voice of a teacher in their software: listen to them. Go to your users’ schools. Sign in with the front desk. Get a visitor’s badge. Sit in the back of the classroom. Watch students empty out a laptop cart. Watch them use your product. Take notes where your product doesn’t do what they need. You will take a lot of notes. Bring them home with you. Pin them to your wall. Talk about them with your coworkers. Fix them. Go back and visit again.

Then, hire teachers. The role of a teacher is one of the most valuable in all of society, and leaving the classroom is a hard decision, but many make it every year. Many want to stay in education, but don’t know where they fit in the ed-tech ecosystem. This status quo is ludicrous. In an education company, teachers can and should have the loudest possible voice when prioritizing the next steps for your product.

Education is not rocket science — it’s harder. It can’t be solved with technology. It’s solved with a voice and clues from your users. The solutions become obvious when you get outside your office. Count yourself lucky if you find a way to bring those voices back with you.

—

Elijah Mayfield is the founder and CEO of LightSide Labs, dedicated to improving student writing skills. LightSide was acquired by Turnitin; Elijah now serves as VP of New Technologies at Turnitin. He has been recognized as a Global Shaper by the World Economic Forum and has been featured by NPR, Education Week and edSurge.