AI agents complete entire online courses as universities face an unprecedented academic integrity crisis

I caught my students using AI to cheat in an exam -- here's what universities must do to stamp this out

The rapid spread of generative AI tools has exposed deep vulnerabilities in conventional methods of academic assessment. This challenge came into sharp focus for me when I found evidence of widespread unauthorized use of artificial intelligence in an examination for my course on welfare economics and social-choice theory at Brown University in Providence, Rhode Island, last semester. Some students were reluctant to return to the classroom after a campus shooting earlier in the academic year, so I administered the midterm test as a take-home exam for the first time in nearly two decades of teaching the course. Although it was more demanding than previous exams, the average score was 96%, far above the course's historical range of roughly 65-80%. Suspecting AI-enabled cheating, I held the final test of the course as an in-person, closed-book examination. The average score plummeted to 48.6%. Furthermore, several students either dropped the course or did not sit the final exam. The shocking aspect of this episode was not that cheating seems to have occurred on the take-home exam, but how prevalent it appeared to be. Universities have struggled to respond adequately to this challenge. But they cannot afford to ignore it -- the very credibility of learning and assessment is at stake. Two things require immediate attention: restoring a culture of academic integrity and adapting student learning and assessment methods for the age of AI. The prevalence of cheating at universities is sustained, at least in part, by the reluctance of academic leaders to confront it decisively. When university leaders fail to defend academic integrity unequivocally, they help to create a culture in which cheating is normalized. My own experience and the countless messages I have received from faculty members at other institutions suggest that administrators often respond to evidence of extensive cheating with hesitation or denial. In some cases, this seems to stem from a desire to protect the institution's reputation and 'market value'. Yet, it misunderstands the factors that ultimately give a university its standing: high standards, a strong work ethic and genuine academic excellence.

A.I. Agents Are Taking Entire Online Courses for Cheating Students

As colleges and students embrace virtual classes, the ease of A.I. cheating is raising questions about the value of an online degree. 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.

A.I. Cheating Threatens the Value of Online Degrees

As colleges and students embrace virtual classes, the ease of A.I. cheating is raising questions about the value of an online degree. 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 it 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.

AI agents are sitting through students' online courses, and colleges are struggling to stop them

The technology has moved from helping with homework to effectively impersonating students inside virtual classrooms. AI cheating has already gone from asking a chatbot to write an essay to using tools that can make that AI-written work look human. We've previously talked about AI tools designed to help students cheat and the growing problem of AI detectors struggling to keep up, as well as how commonly students are already using AI for homework. Now the technology isn't merely helping students do the work. In some cases, it can effectively be the student. Just tell the AI to take the class According to a New York Times report by Dana Goldstein and Alan Blinder, students can take things a step further and hand an entire online course over to an AI agent. As per the report, AI agents can be given simple instructions such as "log in and complete my quiz" and then operate inside online learning platforms. They can watch prerecorded lectures, take tests, write papers, and even participate in discussions with classmates. That's particularly problematic because online education is already built around digital, often asynchronous coursework. Students can complete lectures, assignments, and tests remotely, sometimes without ever interacting directly with their professors or classmates. More than half of American college students took at least one online class last year, according to the report. Recommended Videos And the AI doesn't necessarily have to announce itself. The Times tested leading AI tools including ChatGPT, Gemini and Grammarly and found that none refused requests to write papers on a student's behalf. AI companies also haven't prevented agents from logging into platforms such as Canvas, Brightspace and Blackboard, although those services are researching ways to detect agent use. The problem becomes even harder to police because these agents aren't necessarily limited to one-off assignments. The online learning platforms themselves can look for clues such as how long a student spends on a task, but the NYT reports that there is currently no fail-safe way to block agentic cheating. Some educators have even asked AI companies to make their agents identify themselves when operating inside online courses, but those companies have resisted, citing privacy and security concerns. The problem is bigger than catching cheaters That leaves universities with a particularly awkward problem. AI detection tools can make mistakes, making it difficult for professors to prove that a student cheated, while administrators may be reluctant to punish students without concrete evidence. The result is an inconsistent response, with some professors introducing in-person exams while others simply allow AI use. Some educators are now redesigning online courses around work that's harder for AI to fake, such as personal reflections and discussions about what students found confusing or interesting. Others are bringing certain exams back into physical classrooms. But that creates a much bigger question about what online education actually means when an AI agent can attend the class instead of the person paying for it. The problem is no longer whether AI can help students cheat on an assignment. It's whether a student can complete an entire degree without actually doing much of the learning. And for colleges that have spent years arguing that online education can make higher education more accessible, that's a particularly uncomfortable problem to have.

College Kids Are Using AI Agents to "Take" Entire Online Classes for Them by Logging Directly Into the Course Management Software

Can't-miss innovations from the bleeding edge of science and tech With the rise of agentic models, why stop at using chatbots to look up answers? Some students are directing AI agents to "take" entire online courses for them, letting the software run in the background while they focus on better uses of their time, like doomscrolling on their phone or gaming. The New York Times unpacked the trend in a recent story, and it's a testament to how extreme AI-enabled cheating has become, but also of the perils of getting an education in a virtual environment instead of a physical one. Professors can't rely on supervised exams and oral presentations in online courses like they're starting to shift to in-person ones. Since everything is behind the screen, students can ask an AI agent to log into learning platforms like Canvas and Blackboard and complete quizzes for then, watch lecture videos, write papers, and even pose as the student and participate in discussions, a type of assignment in which they're expected to chat with classmates and sometimes the professor about assigned reading. AI companies are doing little to stop this. The NYT found that none of the three most popular chatbots with students -- ChatGPT, Gemini, and Grammarly -- would refuse requests to write an entire paper on a student's behalf. And they also don't stop their models from logging into platforms like Canvas and Blackboard, and have ignored the request from educators that their models should identify themselves while they're used on there. The reasoning can sound quite specious. Perplexity, the company behind an AI web browser, told the NYT in a statement that requiring AI agents to identify themselves on learning platforms "would put the student's privacy and security at risk." Professors feel powerless to stop the trend. Often, their university may have a partnership or deal with an AI company that provides access to a custom version of its models. This sends a mixed message to students: is AI encouraged, or verboten? Some professors will turn a blind eye AI use, and others will heavily punish it, but as a whole, it's every instructor for themselves. "Am I just interacting with bots here?" Karen Costa, who's spent two decades teaching line courses, recalled asking herself, per the NYT. "I'm not getting any guidance on how to dissuade their use," added Carol Sewell at the California State University system, which signed a multimillion dollar deal with OpenAI last year. "I'm getting opportunities to learn more about using it." Despite the clear potential for teaching -- and negative effects they cause, such as impaired critical thinking and memory loss -- many universities are reluctant to proscribe AI entirely. That doesn't mean some haven't been trying to find ways around it. The University of Chicago Law School recently banned using phones, tablets, and laptops in class for freshman-level courses, as part of its new "AI strategy." Princeton University dropped its over century-old Honor Code tradition which let students take their exams without supervision, after an AI cheating scandal rattled the Ivy League institution.

The State of Online Courses: From AI-enabled Teaching to AI-enabled Cheating

The question that academics are now asking is whether an online degree should be given the same value as one obtained in a campus When agentic AI came calling, every technology expert worth their chip exclaimed that the future was here and fast. However, it looks like colleges and professors are now struggling to counter largescale cheating by students. Students are not just using chatbots to look up answers, they are even directing AI agents to take up the online courses on their behalf. Guess future Ranchod Das Chanchads of the AI era would not be requiring a "Chote" a.k.a. Phunsukh Wangdoo to attend college as a proxy! A report in The New York Times details the latest trend and recognises the fact that AI-enabled cheating is no more an exception but could soon become the rule. More so, when students seek education within a virtual environment as against a physical one. Imagine how teachers would miss supervised examinations or oral discussions in a digital classroom? Cheating has quickly gone beyond mere copy-pasting of chatbot-created essays and problem sets, the report says while noting that AI agents now let students outsource a semester of work. With simple commands like "log in and complete my quiz", the student can let it loose as an imitation of themselves. What's more, they do not identify themselves to the teacher. Given that everything in such a classroom happens behind the laptop or desktop screen, students are finding the easier way out - they simply ask an AI agent to log into learning platforms and complete the quizzes for them, watch lecture videos and write papers. All this is done by posing as the student in the virtual classroom while the real one takes the chill pill somewhere. The NYT report says some AI agents even pose as the student and participate in discussions that are part of classwork delegated by the professors. In such assignments, the student is expected to discuss with classmates and sometimes with the teacher about the assigned reading material and questions thereof. Per the report, companies delivering AI solutions couldn't care less about this trend. The three popular chatbots - ChatGPT, Gemini and Grammarly - never refused requests from users to write an entire exam paper on behalf of the student. And the models do nothing to stop the agent from logging into platforms like Canvas and Blackboard with impunity. "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," the article says. Most chatbot owners have ignored requests from schools and colleges asking the models to identify themselves while they are logged in as a proxy for the student. And the reasons are as crazy as they come. The NYT revealed that Perplexity said in a statement that such a step "would put the student's privacy and security at risk." Academics are finding it difficult to measure the extent to which this malaise has caught on, given that while some universities are completely online, many of the traditional schools are expanding their online offerings. Students are now free to earn full degrees online, while in some cases they take a limited number of classes individually. The teaching faculty say they are largely powerless to curb this trend, given that universities may have partnerships with an AI company providing a custom version of its models. As for students, they face a dilemma in that while the school itself is encouraging AI use, they feel compelled to use it once a peer in the class does so - after all, FOMO is a reality in the modern world. Several academics to whom the NYT spoke with concur that things are getting worse as teachers often wonder if they're interacting with bots or actual humans. Others wonder if there could be guidelines laid out on how to dissuade AI use for cheating. If so, who would lay them down? The university or the chatbot-maker with whom they've done a multi-million dollar deal? Some colleges are taking early steps to control AI use in the wake of research papers suggesting impaired critical thinking and memory loss. The University of Chicago Law School recently banned the use of phones, tablets and laptops in class for freshman courses. Princeton dropped its century-old Honour Code tradition that let students take exams without supervision. All of which leads us to an important question - Should degrees from online programs have equal value to those earned on campus? However, policymakers and academics cannot instantly veer away from their long-term support for digital learning as a tool to generate social equality among students, especially for those who cannot afford to educate their wards. For these children, online classes put degrees within their reach. The problem though is whether AI accessibility is causing serious concerns about this format, especially in higher education. And the dilemma is visible among the teachers who seem to be using their own ways to tackle it. Some just shrug their shoulders and let it happen, others hand out poor grades while the rest allow unbridled use of AI because they know bots can seldom go wrong with their answers.