Artificial intelligence threatens to undermine online education, a growing business for colleges and universities, as professors struggle to counter a groundswell of tech-enabled cheating.

In physical classrooms, many educators have shifted to oral presentations and handwritten exams to try to prevent misuse of an unregulated technology. But those strategies can be harder to execute online, where more than half of American college students took at least one class last year, up from about a third in 2019.

In many online courses, whether at community colleges or globally renowned universities, students may never meet with their professors or their peers, even over video. Those classes are self-paced with prerecorded lectures, also known as asynchronous. Assignments and tests are submitted digitally.

While some universities are completely online, an increasing number of brick-and-mortar schools are expanding their online offerings. Sometimes students can earn full degrees online. In other cases, they take individual classes.

In this realm, A.I. cheating is quickly accelerating beyond the copy-and-pasting of chatbot-produced essays and problem sets. A.I. agents have made it possible for students to outsource a semester of work. With simple commands — “log in and complete my quiz” — a student can unleash what is, essentially, an imitation of themselves. Agents can watch lecture videos, take tests, write papers and chat with classmates about reading assignments.

There have always been questions about whether degrees from online programs are of equal value to those earned on campuses. But policymakers and education leaders have spent two decades championing online learning as an equity tool allowing nontraditional college students, like low-income working parents, to study wherever and whenever is convenient for them. Many students say online classes put degrees within reach.

Now, though, the accessibility of A.I. is deepening misgivings about the most accessible form of higher education — just as many students are struggling with the academic aftermath of the pandemic and questioning whether they have the skills to succeed in college.

In interviews, instructors from coast to coast described a sinking realization that A.I. had sneaked into their digital classrooms. Karen Costa, who has spent close to 20 years teaching online courses, initially reacted with eye-rolls when people would talk about how A.I. could intrude on higher education.

But last year, she said, she sensed an “influx” of A.I.-generated work from her students.

“Am I just interacting with bots here?” Ms. Costa, who herself earned an advanced graduate study certificate through a primarily asynchronous program at Northeastern University, recalled thinking.

Kimber Peters, who teaches online for Southern University, sometimes notices that submissions carry clues — like formal essays with the common A.I. tick of using the pronoun “you” — that suggest a chatbot was behind the writing.

Edward Gray, who has led online courses for various universities for decades, has another tell: when a student submits “a very polished paper that references all of these resources that were not required.”

Kathryn Kysar has taught community college since the 1980s. Lately, she has seen students perform abysmally on an initial assessment of punctuation skills but soon after submit sterling work studded with colons, semicolons and dashes.

“It’s just so obvious,” said Ms. Kysar, who has facilitated online courses for about 15 years and blames regulatory and leadership failures for higher education’s troubles with A.I. misuse. “Those of us who have been teaching for a long time and are editors and writers can usually recognize the cheaply written A.I. stuff within 30 seconds.”

She now requires in-person midterm and final exams for some classes, with a paper dictionary on hand at the front of the classroom.

Obvious Cheating. No Obvious Solutions

Students across the country often ferociously deny any wrongdoing and administrators, faculty members say, are regularly reluctant to impose discipline. In part, this is because A.I. detection software can make mistakes, so there is no concrete proof of misconduct. And some administrators, professors fear, do not want to risk losing a paying customer in an era of tightening budgets and shifting demographics.

Education leaders are also eager to embrace a technology that could reshape the working world.

The result is an uneven response to A.I. misuse, with some professors shrugging and others handing out poor marks. Still others allow unfettered access to A.I.

Many professors feel as if they are on their own.

“I’m not getting any guidance on how to dissuade their use,” said Carol Sewell, who teaches in the California State University system, which has eagerly embraced A.I. “I’m getting opportunities to learn more about using it.”

Ms. Sewell, who teaches gerontology, doubts that much of the cheating is “fundamental dishonesty” and likely reflects the frenzied lives of many students.

Still, she said, “I’m from a different generation — I’m the example of who they’re learning about in my classes — so I think they think they can fool me.”

Some faculty members have called for government regulation of A.I. tools used by students. So far, acting on their own, tech companies have done little to dissuade A.I. cheating. The problem affects both traditional and online classes, but is harder to police when students are completing assignments and tests from home, far from professors’ watchful eyes.

A.I. Platforms That Don’t Say No

In tests by The New York Times, none of the three leading A.I. tools among students — ChatGPT, Gemini and Grammarly — responded “no” to prompts asking for chatbots to write a paper on a student’s behalf. Versions of the software licensed by individual schools can impose greater barriers, but nothing prevents students from getting around those limits with personal email addresses or second devices.

And A.I. companies have not prevented agents from logging into learning platforms like Canvas, Brightspace and Blackboard, which are used by colleges to manage online classes.

Those platforms can look for signals of agent use, such as time on task. But several said there is no fail-safe way to block agentic cheating, though they are researching it.

“Our customers are making it clear that there has never been more pressure on academic integrity,” said Jason Weaver, Blackboard’s chief of staff.

The company recently acquired an app that learns student typing patterns to authenticate their work.

Some educators have asked that A.I. agents working inside online courses identify themselves so professors can monitor how they are being used. But A.I. firms have so far resisted such transparency.

Perplexity, a company that makes an A.I.-powered web browser popular with students, said in a statement that limiting users’ ability to use agents inside online learning systems, or requiring those agents to identify themselves, “would put the student’s privacy and security at risk.”

Neither Google, which makes the Gemini chatbot, or OpenAI, which makes ChatGPT, responded to questions about their agents posing as students. In written statements, each of the rival companies said they were working alongside educators to protect meaningful learning.

(The New York Times is suing OpenAI and Perplexity for copyright infringement.)

Bringing Online Students Offline

Experts have warned that even more novel forms of A.I. cheating are emerging. Wearable devices, like glasses, can allow students to request help even while working in a secure web browser. A.I. avatars can impersonate students over video.

In the absence of technical solutions, some educators, like Ms. Kysar, have begun infusing more real-world elements into online courses.

“We don’t want to take away access,” said Marina Aminy, executive director of California Virtual Campus, which provides online community college courses. But she said professors could add flexible, in-person elements, like a classroom-based exam offered on multiple days.

Some institutions are also changing how online assignments are designed, leaning on personal responses that are more difficult for A.I. to fake.

In a recent online statistics class from the School of Professional Studies at the City University of New York, the participation grade was based, in part, on a discussion forum in which students were required to name the “clearest” and “muddiest” concepts for them from each unit, and why.

Since students were encouraged to surface their struggles, there was little incentive to cheat, said Ruru Rusmin, an assistant dean and expert in educational technology.

Jordan Canzonetta, an associate professor at Lewis University in Illinois, said she emphasizes the pitfalls of relying on A.I. — inaccurate information, reduced skills — while encouraging students to turn in their own work.

“My biggest struggle is getting students to trust that they do have writing ability, that they have important things to say,” she said.