Imagine you are a college president and believe two things. One, an important goal of college is to increase students’ chances of achieving the future they want. Two, artificial intelligence (AI) will change large parts of the economy in the coming years and decades. Given those beliefs, what should you do?

First, you should encourage everyone at the college to use AI more. Although the details of future careers are hazy, there can be no doubt that students who know how to use AI well will, on average, outperform those who do not. The best way to improve at AI is through practice. Moreover, while your primary focus is on students, encouraging faculty and staff to use AI as much as possible can only help. As part of this effort, you should provide institutional licenses that give everyone access to the professional versions of a range of AI tools. You should also favor applicants in admissions and hiring decisions who demonstrate a strong interest in AI.

Second, work with faculty who are already enthusiastic about AI to make their courses as AI-friendly as possible. As a college president, you know better than anyone how difficult it is to shift faculty opinion—especially on academic matters. Rather than trying to convert skeptics at the outset, begin with those who are already open to experimentation. They can demonstrate what responsible, effective AI integration looks like in practice. Courses taught by such instructors can model responsible AI use in student work. The more exposure students have to AI in these settings, the more competent they will become.

Third, allow students to use AI for all work completed outside the classroom. Recalcitrant faculty will howl about this intrusion into their domain, but a general rule is the only effective way to deal with such holdouts—and, in any case, students are already using AI for most out-of-class work anyway. Next, require every course to culminate in a public final project created with AI support. Preventing faculty from using X—whether a midterm exam or an in-class writing exercise—is a losing battle; requiring everyone to include Y is far simpler.

Taken together—unrestricted AI use outside the classroom and a mandatory AI-assisted final project—these two policies would ensure that students graduate not merely familiar with AI but fluent in its use. Faculty will object, of course. Here are the most common complaints—and how you should answer them.

Students need to learn to write essays without AI.

If you believe that learning how to write an essay or solve an integral without AI is important, then practice that skill in the classroom. You run your classroom. However, the college has decided that our students will be far more proficient in using AI than graduates of other institutions. Outside your classroom, they will use AI as much as possible.

There are no final projects for my course that could reasonably use AI.

Yes, there are. Pick any topic in your field. Have your students create a video, a website, or an interactive game—anything that educates people about the topic. Of course, you will need to learn about AI tools, if only so you understand how students might tackle these projects and how much time they will take.

AI projects are better suited to later courses. There is so much introductory material to cover in the first two years of our majors that we can’t afford to divert time from our traditional lecture-and-exam approach.

The college has identified AI as its key strategic priority for the next 20 years. Students need to practice, and final projects are the perfect occasion for that practice. Students also need a professional portfolio of projects to demonstrate their skills to potential employers. Even better, you will soon discover that students work much harder on public projects than they do on projects that only you will see. We are not claiming that what you used to spend time on is unimportant. The college has just decided that AI is more important.

I don’t know enough about AI to teach my students how to use it.

Time to learn! And we are here to help you.

Thoughtful college presidents think in terms of causal effects. If a student attends my college, she or he will have access to a range of internships/jobs/career opportunities. If she or he had attended a different school, she or he would have secured different internships/jobs/career paths. The causal effect of attending my college is the difference between those two outcomes. To the extent I believe AI will fundamentally change the economy over the next few decades, the best way to ensure a positive causal effect of attending my college is to ensure that each student is an expert in using AI.