U.S. education is in crisis. Poorly paid teachers, funding tied to property taxes, and the lack of a national curriculum have produced a fractured national landscape for primary education. The ongoing coronavirus pandemic has exacerbated that situation, forcing unprepared teachers to teach remotely.

In higher education, enrollments were dropping as tuition climbed even before the ongoing pandemic. COVID-19 has only accelerated that trend.

The country needs a new paradigm. Artificial intelligence is part of the answer.

Man has known since antiquity that the best education is delivered one-to-one by an experienced educator. But that is process is expensive and labor-intensive and cannot scale. The result is the classroom-based instruction that, despite its obvious drawbacks, has persisted for centuries.

Forty years ago, there was excitement about so-called adaptive learning based on expert systems, elaborate “if-then” decision trees that led students on a path through content compiled by subject experts. But building the knowledge bases required for such systems was time consuming and costly, and ultimately couldn’t scale.

A.I. has come a long way since the early adaptive-learning days. Researchers are beginning to understand how the brain learns and are applying that understanding to networks of algorithms that learn on their own. Those methods, in turn, are being used to build software that makes it easier and more engaging for students to study. These A.I. tutors, software systems that students interact with online, can give everyone individualized attention. Suddenly, one-on-one teaching can scale.

A.I. tutors are eternally patient. What’s more, their teaching is not based on guesswork or intuition, but on data. Learning algorithms uncover patterns about how individual students perform, and companion algorithms optimize teaching strategies accordingly.

Kidaptive, a startup out of Stanford University, has an adaptive-learning platform for children that paints a vivid picture of the learner’s skills, interests, and activity across products and over time. That information is used to personalize education with real-time assessment, such as interactive teaching through animated story lines delivered on iPads or other touch-screen devices.

My own company, Riiid, has prepared over a million students for English proficiency exams on our A.I. tutoring app, Santa. We are now developing A.I.-driven platforms for education companies, school districts, and national governments in the U.S., South America, the Middle East, and Asia. Other companies are also preparing to launch A.I. tutoring systems in the coming months.

Studies show that these systems can raise student performance well beyond the level of conventional classes and even beyond the level achieved by students who receive instruction from human tutors.

Advances in generative adversarial neural networks and A.I.-driven animation, meanwhile, are creating realistic avatars that will soon put a friendly face on A.I. tutors. Smart speakers, millions of which are already in people’s homes, will also deliver classes powered by A.I. This is not fantasy or science fiction. It is just getting started, but it is happening now.

Teachers will not disappear, but much of the drudge work of teaching will. A.I.-powered apps are not only taking over grading, the bugbear of every instructor, but are collecting meaningful data that would otherwise disappear.

Bakpax is an A.I. platform that uses computer vision to read student homework and grade it automatically. Gradescope assists teachers by converting schoolwork to digital form with tools to speed grading. Both systems collect data on student performance that would otherwise end up in the trash can. That data can, in turn, be analyzed by A.I. systems to uncover useful patterns, which teachers can use to adjust their methods or syllabi, forever optimizing their own performance.

Most importantly, A.I. tutors are a step toward democratizing education. Eventually, any parent with an Internet connection will be able to give his or her child an effective, personalized learning experience regardless of zip code.

A.I. can help, too, in shaping higher-education curricula to maximize the successful employment of graduates, which is itself a critical metric in attracting students and justifying high tuition fees.

Never has education needed these solutions more than now. With COVID-19 still going strong and a rise in infections projected this fall, instruction will continue to be largely online.

Top university courses can now be taken online for free or a fraction of the cost and commitment of a four-year degree. Project portfolios are replacing academic credentials as the most important metric for many prospective employers. In recent years, many schools turned to full-fee paying international students to shore up fragile finances, but the pandemic has cut that short.

Analysts are predicting a wave of consolidation as smaller schools merge or close altogether. America’s roughly 4,000 colleges and universities may shrink by as much as half that number within the next decade.

Education administrators need to explore available A.I.-powered solutions to deliver quality education while protecting faculty and students alike; streamline curricula and align educational offerings with government and private sector needs; and optimize budgets, building solid long-term plans for financial survival. Parents need to pay attention.

Education is the foundation of a healthy economy. Equity in education is crucial for a cohesive society. The tools exist to fix the U.S. education system. The only thing missing is a plan.

YJ Jang is CEO and founder of Riiid, an A.I. education company that builds systems to power online and offline learning.