Will AI make American education even more unequal?

The proverbial genie is out of the bottle. Generative artificial intelligence (AI) is being developed and used at a dramatic pace, much faster than the ability of policymakers to understand, much less regulate, its use. AI technologies are changing the way services are provided and businesses operate, and it is already clear that this new technology will produce winners and losers. The technology has the potential to transform almost every sector of society.

In education, the changes produced by the use of AI are particularly dramatic and unsettling. Generative AI tools such as ChatGPT are now widely used by many students in K-12 and higher education. Undoubtedly, professors will also use these tools to publish “their” work.

Widespread use is leading to profound new questions related to intellectual property rights, authorship and the need for new rules to guide attribution. In the absence of clear rules, some universities and school districts have banned the use of generative AI altogether. However, given widespread access to AI tools, it’s unclear whether such bans are even enforceable. 

Rather than simply attempting to ban or limit the use of AI, it is important for policymakers to focus on a few critical questions. Who will have access to cutting-edge technology as it’s developed? Access will be a key factor in determining whether AI will be used to further educational opportunity or reduce it. Similarly, as the pace of change accelerates, how should educational institutions set rules for the use of AI by students and faculty? Most of all, will the changes brought about by AI exacerbate existing inequities, or can AI be used to ameliorate them?

This is not a trivial matter. As we’ve already seen with other technological breakthroughs, education is typically the caboose on the engine of progress. Innovations — such as digital libraries, cloud storage and the use of tablets — were widely used in industry years before they became available to students in America’s classrooms.

Of course, being in the caboose is not always a bad thing, especially if those at the head of the line are the first to make costly mistakes. For example, we’ve recently seen some technological changes implemented without adequate training provided to the users, or with insufficient consideration of how new products will interact with older ones that are still in use (Remember Apple’s Newton?). Disruption may be a good thing — unless it leads to work stoppages, paralysis and greater inefficiency.

I’m a member of the U.S. Department of Education’s new committee on the use of AI in education. In that role, I intend to help the administration devise policies that will make it possible to utilize the extraordinary potential of AI to advance educational opportunities for all.

I’m open to suggestions, especially from the tech sector. As we’ve seen with other promising innovations in technology, too often breakthroughs accelerate existing inequities rather than ameliorate them.

During the pandemic, for example, virtual learning became essential for millions of students as schools closed, in some cases for nearly two years. A 2020 Pew research poll found that one in five parents with homebound school-age children said that their kids were not able to complete school work because they didn’t have a computer at home. Shockingly, nearly one-third of parents in the sample reported that their children had to do their homework on a cell phone. 

Access to quality education continues to be a fundamental challenge in the U.S., despite the country’s wealth and expertise. AI could help to address these inequities if policymakers ensure that this is a national priority.

In some schools, AI is already being used to personalize learning, provide real-time feedback to students and empower teachers and students by providing access to even more information than they have now. Such interventions can help in reducing academic disparities, which tend to follow predictable patterns with respect to the race and socioeconomic backgrounds of kids. 

The administration and Congress must enact regulations requiring tech firms to invest in the development of generative AI that helps kids and schools. If they don’t, the firms are more likely to simply treat education as a market to be exploited — which would likely further the existing inequities in education for generations to come.

We are in a period of rapid and dramatic change. We have an opportunity to use this moment to influence educational trends in ways that make the future more equitable, just and sustainable. Let’s not squander this opportunity.

Pedro A. Noguera dean of the Rossier School of Education at USC.