Breakthroughs in research in fusion energy and in education have at least one thing in common: Over the last few decades, pundits have often said big breakthroughs will soon take place.

Just as with the need to find cheap clean energy, the need to find effective ways to improve our education system is growing. This has been crystallized by the well-documented learning losses associated with COVID-19 and by the work of analysts who have translated that learning loss into estimates of lost earnings of $1.6 trillion dollars nationwide. But make no mistake, these COVID-induced learning losses are exacerbating long term trends, where American students, especially the lowest performing ones, have been falling further and further behind before COVID-19.

While I don’t expect a fusion-level breakthrough this year, I do expect some impressive breakthroughs in education research and development by year’s end. Here are some of the big trends in education R&D that might bring us closer to breaking the pattern of failed prognostications, producing the conditions for some big bangs.

AI advancements driving investments

Just in case you were occupied with important things (like family) during the holiday season, you might have missed the release of the OpenAI GPT-3 chatbot. While the jury is still out on how much this chatbot will accomplish, there is no doubt it represents the coming together of decades of work on artificial intelligence and machine learning.

Right now, the chatbot can produce essays and other reports I would say are about C grade level work in a good university. But the 3 in GPT-3 tells us everything we need to know about the future. Indeed, GPT-4 is already in use on a limited research-oriented basis, but there is no question the coming together of artificial intelligence, machine learning, and large language modeling will drive investments by science agencies, foundations and commercial enterprises in the coming months and years. And indeed, companies such as Microsoft are investing heavily in OpenAI GPT.

If education research reaches anything even close to the fusion breakthrough, it may be in this field.

Balancing privacy with opportunity in big data

We have been talking about big data for years — indeed, Georgetown University has gone so far as to leave big data behind for its “Massive Data Institute.” While there has been nothing as dramatic as the Open AI chatbot, we are starting to harness faster, cheaper, and more powerful computing power and statistical models to milk insights out of the sea of data that surrounds us.

One of big data’s defining characteristics is that much of it is generated as a byproduct of other processes. For example, schools routinely gather information about student attendance, but those data are now being merged with data from health, criminal justice and welfare data systems to provide deeper insights into why students may be chronically absent. These “administrative data systems” are often designed for one purpose (fiscal reporting in the above example) but are merged with other data systems to yield insights into, e.g., student performance previously inaccessible.

The big breakthroughs here will require, first, a more liberal interpretation of the growing body of laws (such as the 2018 Foundations for Evidence-Based Policymaking Act) that require merging federal data sets. And second, a strong solution to the heightened privacy concerns that flow from merging disparate data sets, each of which may have strong privacy protection, but when merged can produce expanded opportunities for identifying individuals. 

Speeding up research

Consider the miracle of the COVID vaccines. Years of work on mRNA preceded the outbreak of COVID-19 and laid the foundation for the breakthroughs that produced effective vaccines in 10 months instead of 10 years. But the way in which the federal government spurred the translation from that laboratory work to the rollout of effective vaccines was not ordained.

The federal government preordered large numbers of vaccines from multiple producers to increase the odds one or more vaccines would come to market. But education research and development operates not at warp speed, but at, well, the speed of government.

Imagine if the government had bet on Moderna alone and worked at normal speed. After some number of years, if Moderna didn’t work, the investment would shift to AstraZeneca. If that didn’t work, it would shift to Johnson & Johnson. And many more of us would be dead.

Of course, that is a stylized version of what would happen in pharmaceuticals, but it is NOT a stylized version of what happens with education research. Most education research is conducted by academics in universities or large contract research shops, and most is “field initiated”— that is, it is generated by those researchers, many of whom lack either the incentives or the skills to take positive research results through product development and widespread adoption.

This year’s federal budget gave the Institute of Education Sciences (the agency I head) a down payment of over $30 million to facilitate the movement of education research into a more modern, rapid cycle mode. Private philanthropies are also looking at ways of speeding up education research and development to uncover more things that work and to translate positive findings into products that improve learning outcomes. With a bit of luck, modernizing the education R&D infrastructure will lay the foundation for breakthroughs in the years to come.

These are the three biggest trends I will be watching, and they will likely combine in ways that we cannot yet foresee.

But here is one outcome I think will happen as these trends develop: Education will move faster and further from the “factory model” our schools have relied on for the last 100 years to a system of education that is more personalized, using data and new analytics to tailor education to fit learner needs and improve outcomes.  

Here’s looking at you, 2023 — a year of big breakthroughs for education!