AI Cheating, Leaked Papers, Marking Errors: How Exam Protests Went Global

What might have been a summer of quiet family anxiety has erupted, in several countries at once, into public unrest. On August 16, 2026, The Guardian published a sweeping account of a phenomenon that education researchers are still struggling to name: exam season has gone global — not in reach, which it always had, but in volatility. Mass student protests in Mexico. Portugal’s worst education crisis in decades. And in India, a paper-leak scandal that affected millions of students, was linked to more than a dozen student suicides, forced the education minister to resign, and grew into one of the most serious political challenges Narendra Modi’s government has faced in years.

At the heart of it, experts say, are two forces colliding: the explosive growth of AI-assisted cheating, and institutions’ rushed attempts to digitise and automate examinations — all against the backdrop of a precarious global economy where a single test score can decide a young person’s entire future.

India: The Paper Leak That Shook a Government

The largest and most ground-shaking incident was India’s. An exam paper leak affected millions of students and was linked to more than a dozen taking their own lives. The scandal forced the education minister to resign and challenged the government’s hold over its 1.4 billion citizens.

The protests themselves took an unlikely form. Students from across India joined mass demonstrations led by the Cockroach Janta party — a youth-led protest campaign that began as an online satirical joke and grew into a nationwide outcry. What started as anger over exam irregularities widened into a broader movement about unemployment and opportunities for young people in a country where government job examinations are among the most competitive in the world.

Central to the story is what researchers call the “education mafia”: illicit commercial industries that exploit student desperation by stealing and selling exam papers, something that has become widespread in India and elsewhere. When a single test determines a person’s entire future, cheating stops being an individual moral failure and becomes a system-level industry.

Mexico and Portugal: Two Kinds of Institutional Failure

Mexico’s crisis came from suspiciously good results. The National Autonomous University of Mexico (Unam), the country’s most prestigious university, noticed abnormally high marks in its entrance exam results — a pattern consistent with widespread cheating. The response: approximately 58,000 students were forced to retake the exam, sparking protests from applicants who had done nothing wrong.

Portugal’s crisis came from automation itself. Education minister Fernando Alexandre pushed to fully digitise the exam marking process, and the result was serious marking errors that landed the government in the country’s worst education crisis in decades. Parents, teachers, and students protested, demanding Alexandre’s resignation. Where Mexico’s failure was a failure of exam security, Portugal’s was a failure of the technology that was supposed to fix everything — a warning for every institution betting that digitisation alone can restore trust.

“It’s Spy Technology, Really”

The connective tissue between these crises is how thoroughly AI has changed the economics of cheating. Dr Thomas Lancaster, a computer scientist at Imperial College London who researches the educational use of generative AI, told The Guardian that cheating is no longer about notes written on a sleeve — it’s sophisticated communication.

“The number of notes you sneak in, they’re only useful if they’re the right notes and if you can interpret them very quickly,” Lancaster said. “But with any kind of communication technology, you can have that expertise waiting outside for you in a human form. They either know the answers; they’ve got the expertise; or they can look them up online. You can go further than that now and have access to AI to communicate directly during the exam as well. It’s spy technology, really.”

Students now use earpieces and miniature cameras, with accomplices — human or AI — on the other end. The NYT reported on August 10 that AI agents are now taking entire online courses: watching lecture videos, sitting tests, writing papers, and even chatting with classmates about reading assignments. The take-home exam, the essay, and the online course — the assessment formats of the last two decades — are all structurally compromised.

Denmark’s Answer: Kill the Essay

The policy responses are diverging fast. Denmark announced this month that from the new term, teenagers will have to verbally defend their written essays — a return to the medieval viva voce as an AI defence. It’s a striking admission: the written essay, a 500-year-old technology for demonstrating thought, is no longer a viable proof of learning.

Dr Sarah Eaton of the University of Calgary’s Werklund School of Education, a leading expert on academic ethics, has implemented similar reforms in her own classes. “AI can generate a perfect essay in seconds. So an essay is no longer a viable way to demonstrate students’ learning,” she said. “We should let it die with dignity and find other ways to assess student learning.”

Eaton’s diagnosis goes deeper than technology. “When you get large-scale fraud combined with systemic failure, there’s almost a guarantee of a breakdown of public trust,” she said. “And once that trust is broken, it’s almost impossible to regain. I can’t imagine how India is going to deal with this, just given the levels of corruption and fraud that we’ve seen.”

She also warns against the quick technological fix — the hope that technology can be “easily and cheaply used to prevent cheating en masse” — when the real failure has been not training educators. Cheating is a human problem, she argues, and assessment has historically been a human responsibility that cannot simply be off-loaded to the latest tech. Mass examinations at scale, she says, “dehumanises the learning process and the assessment process,” and perpetuates the idea that examinations — rather than learning itself — are the point.

The Uncomfortable Pattern

Step back and the pattern is uncomfortable. AI has made generating perfect coursework trivially cheap, so institutions are digitising and scaling assessment to compensate — which introduces new failure modes like Portugal’s marking errors, which destroys public trust, which fuels protests, which pressures governments into crisis management rather than reform. The countries hit hardest are those where exams carry the highest stakes: India’s government job exams, Mexico’s university admissions, Portugal’s national school-leaving certificates.

The economic substrate matters as much as the technology. “The results of one examination determining a person’s entire future really creates untenable pressure on young people,” said Eaton. “This is a tragedy when young people are dying by suicide because of an examination outcome.” In a tough global job market, that pressure is rising, not falling — and no proctoring software can fix that.

What’s coming is a decade-long renegotiation of what assessment means. Oral defences, in-person exams, project portfolios, and AI-integrated assessment (testing students on their aptitude for working with AI) are all on the table. The one option that’s off the table is the status quo: the take-home essay is already dead. Denmark is simply the first country to say so out loud.