Contract Cheating: The Billion-Dollar Threat That AI Detection Can't Solve

Fifty dollars. That's what it costs to buy a custom-written essay that passes Turnitin, passes GPTZero, matches the rubric, and has never existed before. The essay is original, human-polished, and indistinguishable from honest work by any content analysis tool on the market.

Contract cheating is a billion-dollar industry, and the tools most schools rely on for academic integrity were never designed to catch it.

Scale

The contract cheating industry is estimated to be worth over $1 billion USD annually and continues to grow. Essay mills (commercial services that sell custom-written academic work) operate openly online, advertising on social media, campus flyers, and even through targeted ads that appear when students search for homework help.

The Numbers Keep Getting Worse

A systematic review of research on contract cheating found a historic average self-reported rate of 3.52% of students admitting to the practice. In studies from 2014 onward, that number jumped to 15.7%, representing an estimated 31 million students globally.

These are self-reported figures. Actual rates are almost certainly higher because students have little incentive to admit to cheating, even in anonymous surveys.

COVID-19 Accelerated the Problem

When institutions shifted to remote learning during the pandemic, contract cheating surged. Essay mills saw the absence of in-person supervision as an opportunity and used aggressive marketing to attract students. The lack of face-to-face interaction made it harder for teachers to notice when a submission didn't match a student's known abilities.

Why Turnitin and GPTZero Can't Help

Contract cheating is virtually undetectable by content analysis. The reason is straightforward:

Plagiarism Checkers Can't Help

Turnitin and similar tools compare submissions against databases of existing text. Contract cheating produces original text, custom-written for each client. There's nothing to match against. The essay has never existed before, so it passes every plagiarism check.

AI Detectors Can't Help Either

AI detection tools look for statistical patterns characteristic of machine-generated text. Contract cheating often involves human writers: paid ghostwriters who produce genuinely human-written text. Even when essay mills use AI to speed up production, they typically have human editors polish the output, making it undetectable by current AI tools.

The Perfect Evasion

Contract cheating represents the perfect evasion scenario:

• Original content → passes plagiarism detection
• Human-written or human-polished → passes AI detection
• Custom to the assignment → matches the specific requirements and rubric
• Delivered privately → no digital trail in public databases

The only way to catch contract cheating through content analysis is if the teacher notices a sudden, dramatic shift in writing quality. Even this is unreliable and subjective.

AI Supercharged the Essay Mill Business

ChatGPT and similar tools haven't replaced essay mills. They've made them faster, cheaper, and harder to detect. A 2023 analysis explored how AI enables essay mills to:

• Produce work faster, reducing the time per assignment and increasing volume
• Lower costs because AI-assisted production means lower prices, making cheating accessible to more students
• Customize more effectively since AI can be prompted with specific rubrics, word counts, and style requirements
• Scale operations so that a single operator can handle more clients with AI assistance

Together, contract cheating and AI create something more dangerous than either alone: sophisticated, cheap, and nearly invisible academic fraud.

Governments Are Trying to Legislate It Away

Several countries have tried banning essay mills outright:

• Australia banned advertising and providing academic cheating services in 2020 under the Tertiary Education Quality and Standards Agency Amendment Act
• The United Kingdom criminalized essay mill services as part of the Skills and Post-16 Education Act 2022
• Ireland passed the Qualifications and Quality Assurance Amendment Act 2019, making it an offense to provide or advertise academic cheating services
• Several U.S. states have laws against essay mills, though enforcement has been limited

While legislation is important, enforcement remains challenging. Many essay mills operate from jurisdictions where these laws don't apply, and new services appear faster than regulators can shut them down.

Watch the Writing, Not the Text

If you can't catch contract cheating by analyzing the finished product, the only option left is to verify the process that created it.

What the Writing Process Actually Looks Like

When a student writes an essay themselves, the behavioral record shows:

• Gradual composition: text appearing keystroke by keystroke over minutes or hours
• Natural revisions: going back to fix typos, restructure paragraphs, rephrase sentences
• Thinking pauses: periods of inactivity between sentences or paragraphs where the student is formulating ideas
• Consistent typing patterns: keystroke dynamics matching the student's established baseline
• Progressive development: ideas building on each other, with the logical flow reflecting real-time thinking

When a student submits contract-cheated work, the behavioral record shows something very different:

• Large paste events: blocks of text appearing instantaneously
• Minimal revision: little or no editing of pasted content
• No thinking pauses: the "writing" process consists of copy-paste events, not composition
• Mismatched timing: keystroke patterns (if any) that don't match the student's baseline
• Compressed timeline: a 2,000-word essay "written" in minutes rather than hours

This difference is unambiguous. No matter how polished the ghost-written text is, the process of pasting someone else's work looks nothing like writing it yourself.

Baselines Make Evasion Nearly Impossible

Behavioral verification is strongest when it can compare current writing behavior to a baseline, a reference profile of how each student typically writes. This baseline is established through normal writing activities and becomes more accurate over time.

With a baseline in place, even sophisticated evasion attempts become detectable. A student who normally types at 45 WPM with characteristic pauses and a distinctive digraph pattern can't convincingly fake that pattern while manually retyping contract-cheated work. The behavioral fingerprint is too complex to replicate consciously.

A Multi-Layered Response

No single measure solves contract cheating. Institutions need several things working together:

1. Implement process-based assessment. Use tools that track and verify the writing process, not just the final product. This is the single most effective technical countermeasure.
2. Design harder-to-outsource assignments. Assignments connected to specific classroom discussions, requiring personal reflection, or building iteratively across the semester are harder to contract out.
3. Create a culture of integrity. Students who feel connected to their learning community and understand the purpose of their education are less likely to cheat. The International Center for Academic Integrity (ICAI) provides frameworks for building institutional cultures of integrity.
4. Reduce the pressure to cheat. Many students resort to contract cheating because they're overwhelmed: too many assignments, too little time, too much pressure. Reasonable workloads and supportive academic environments reduce the incentive.
5. Educate students about consequences. Many students don't fully understand the risks: academic penalties, degree revocation, and the fact that they're paying for a credential without the knowledge it's supposed to represent.

The most effective response to contract cheating isn't just better surveillance. It's building educational environments where students value authentic learning. When students can see their own writing process, track their growth over time, and receive meaningful feedback on their actual work, the incentive to cheat diminishes.

We think behavioral verification serves both sides of this equation: teachers get evidence of authentic work, and students get a visible record of their own effort and progress. The goal isn't to catch cheaters. It's to make authentic writing the path of least resistance. Limudi's behavioral verification captures the complete writing process, every keystroke, revision, and paste event, making contract cheating transparently detectable.