AI Contract Cheating 2026: Evidence, Law and Student Risk

Plagiarism-Checker-Online.net Redaktion  |  Updated October 4, 2026

AI contract cheating means outsourcing assessed work and presenting it as your own, including buying a submission produced with AI assistance. A low similarity score does not make that submission legitimate. The academic rules concern authorship and authorized help; the laws against commercial cheating services have their own scope and do not criminalize every student use of AI.

Outsourcing is the issue, not just the technology

TEQSA defines contract cheating as outsourcing assessment to a third party. That may involve a commercial provider or someone the student knows. AI can be part of a provider's workflow, but generating work yourself with an unauthorized chatbot is not identical to hiring a third party. Both situations need assessment under the relevant academic rules.

A provider's promise of original text, human editing, or a detector-safe result does not demonstrate legitimate authorship. Text matching asks whether passages resemble available sources. It does not establish that the named student completed the assessed work. An AI detector also cannot certify that an essay is safe to submit.

What the Reading experiment actually establishes

Scarfe, Watcham, Clarke, and Roesch's study was published in PLOS ONE in 2024. The researchers inserted AI-generated answers into an undergraduate psychology examination system and found that most went unidentified. This was a study of a particular assessment setting, not a survey of essay mill use or proof that every detection system fails.

A finding from one examination setting cannot establish national cheating prevalence, essay mill market growth, or the cost of an investigation. Those questions require their own evidence. Neither a detection experiment nor a student-use survey provides a sound basis for predicting the outcome of an individual submission.

The legal landscape: English institutions

The official explanatory notes to the Skills and Post-16 Education Act 2022 distinguish the definition in section 26 from the provider offense in section 27. The provisions concern relevant assignments for students at English institutions, not a blanket rule for every institution throughout the UK.

Section 27 concerns providing or arranging a relevant service in commercial circumstances. It is a summary offense punishable by a fine; this provision does not prescribe imprisonment. Section 28 separately addresses advertising. Whether a particular service meets the statutory definition requires attention to what it supplies and the statutory conditions; a general-purpose AI tool is not automatically an essay mill.

These provisions target service provision and advertising. They do not create a general student offense of purchasing an essay. That distinction does not authorize submission: a university can investigate outsourced work under its own regulations. Nor does it rule out other legal consequences in a particular case. Consult qualified advice if you face a legal allegation.

Australia has a separate legal framework

TEQSA's legislation guidance explains the 2020 changes concerning academic cheating services in Australian higher education. The framework includes criminal and civil consequences for commercial services and civil consequences for unpaid services. It should not be described as a law introduced in 2023 or as a blanket criminal ban on student AI writing.

Do not transfer the Australian penalties to England, or assume the English law covers other jurisdictions. Provider law and student disciplinary rules answer different questions.

Investigation needs evidence and fair procedure

A review may examine the sources, drafts, account of the writing process, and understanding of the submitted argument. Neither polished prose nor hesitation in an interview proves outsourcing. A student may be nervous, need an adjustment, or have improved their writing. Reviewers should follow the institution's published process rather than treat a single feature as a verdict.

Weber-Wulff and colleagues' detection-tool evaluation identifies reliability limitations under the conditions tested. It does not support a universal detection rate or a guarantee that edited AI text will pass. The EU AI Act student guide also explains why not every educational detector automatically has the same high-risk classification or application date.

What to do instead of buying a submission

If you are behind, contact your instructor, writing center, or student support service about help and extensions. Clarify the boundary between feedback and someone completing the work for you. A service calling itself tutoring does not settle that question.

Keep genuine drafts and research notes, and document permitted assistance accurately. If someone threatens to disclose a purchased assignment unless you pay more, preserve the communications and contact your institution's support team. Do not assume another payment will resolve the problem.

For legitimate writing, our writing workflow guide and policy guide can help you organize your own process. A plagiarism check can support source review, but it cannot predict an investigation, guarantee an outcome, or turn outsourced work into your own.

Frequently Asked Questions

This article is part of our Academic Integrity Guide.