How universities spot AI-generated content

Artificial Intelligence is changing university admissions. How are universities verifying applicants in the age of Chat GPT?By Lydia Suslova Universities around the world are increasingly turning to AI-detection tools and rethinking their admissions policies. So much so that authenticity has become one of the defining admissions trends of 2026, before a potential student even has to think about course essays and projects. Today, the AI debate has moved on to how have universities learned to spot the difference between a student’s own writing and something generated by a machine, and what happens to applicants who get caught? Students applying to universities around the world are falling into a trap they often set themselves. They turn to ChatGPT and other AI tools to draft personal statements because it‘s fast, easy, and practically error-free. Then they get rejected. Not because the essay is badly written, but because it reads exactly like thousands of others. According to a ScholarshipOwl survey of more than 12,000 high school and college students, 97 per cent of Gen Z students have used AI toolsfor schoolwork or university admissions preparation. More than one in five students (21 per cent) admitted using AI to write their college or scholarship application essays; 31 per cent used it for academic essays, 35 per cent for homework, 66 per cent for general studying, and 56 per cent for exam preparation. “Universities are screening applications for AI-generated content and using their own detection software,” says University Application Overview Consultant at the international online school Virtual International School Juliana Bazavluk. “But that’s only one layer. The real work happens when admissions officers read the essays themselves. They go through tens of thousands of applications, and more often than not, they can immediately tell whether they’re reading a student’s real voice or something a machine put together.” According to GradPilot, around 50 per cent of US admissions offices now use some kind of AI to review essays, check for plagiarism, or process transcripts. The actual figure is probably higher, since many institutions don’t openly talk about using these tools. One common red flag is an essay that’s too polished. It might be flawless in grammar and structure, but it still feels hollow, impersonal, generic and devoid of real experience. Another giveaway is the heavy use of clichés and formulaic language. AI-written essays love to talk about leadership, changing the world, personal growth and the importance of technology. These aren’t weak topics in themselves, but AI tends to present them without any concrete examples or genuine personal insight. Admissions officers also keep an eye out for repetitive sentence patterns, unnatural phrasing, and stylistic quirks that are typical of large language models. The UK has taken a particularly strict stance. All undergraduate applications go through UCAS, a centralised online portal that manages nearly all undergraduate applications to universities and colleges in the UK. UCAS requires applicants to confirm they haven’t used AI for any part of their application. Personal statements are reviewed with specialist software, and false declarations can lead to immediate rejection. In many cases, universities don’t offer a detailed explanation: applicants already know they’ve broken the rules. The US approach is somewhat more flexible, though the stakes are just as high. Broadly speaking, institutions have split into two camps. Some have drawn a hard line. Submitting essays written wholly or partly by AI may be treated as application fraud, which can lead to the cancellation of an admission offer. Others have taken a different approach. Duke University, for example, introduced a new essay prompt asking applicants to describe a situation in which they would choose to use or not use artificial intelligence, and to explain their reasoning. The point isn’t to catch applicants out. It’s to get a sense of how they think about technology, ethics and decision-making. Written materials aren‘t the only thing admissions teams are looking at anymore. More and more universities are introducing video submissions and interviews as an extra layer of assessment. They want to see the person behind the application and to check whether the ideas expressed in an essay actually match how that person thinks and speaks in real time. Interviewers are trained to notice when applicants struggle to talk about topics they supposedly wrote about. Students who have leaned too heavily on AI often stumble at this point. Their speech becomes hesitant, vague, or inconsistent with the polished writing in their application. Universities aren’t searching for flawless writing. They’re searching for students. Admissions officers want people who will contribute to academic conversations, get involved in campus life, join research projects, and enrich the university community. A polished but hollow essay can’t prove any of that. But the problem doesn’t start at the admissions stage. More and more students are using AI to complete homework and assignments, and many educators argue that this is quietly undermining the learning process. “This is especially worrying,” says head of Academic Logistics and Quality Assurance at Virtual International School Alice Krylova. “They use AI to do their homework, get strong grades, and then fall apart in the real exam. In the exam hall, they have to think on their own, write by hand, and do it under time pressure.” Educators consistently say the same thing: the responsibility ultimately lies with students and their families. AI can be useful when researching a topic or checking spelling. However, asking a chatbot to write your essay or solve an assignment for you is something else entirely. In the end, it doesn’t fool the university, it only fools the student, who shows up to the admissions process without the skills they need to succeed.