How Teachers Detect AI in GCSE Food Technology Coursework

GCSE Food Technology and Food Preparation and Nutrition coursework has always required students to produce substantial written analysis alongside their practical work. The NEA tasks demand contextual research, food science justification, and detailed evaluations - extended written components that, until recently, were difficult to fabricate convincingly without genuine subject knowledge. The arrival of capable AI writing tools like ChatGPT and Claude has changed that. A student who understands how to construct an effective prompt can now generate plausible Food Technology analysis without demonstrating any real understanding of macronutrients, food provenance, or sensory evaluation.

For teachers running the AQA or WJEC GCSE Food Preparation and Nutrition NEA, this creates a specific challenge. The written components are integral to the qualification - they account for a significant proportion of the NEA marks and are designed to assess whether students can apply food science knowledge to real-world briefs. This guide explains how to use AI detection tools professionally and responsibly when you have concerns about a submission.

What Makes GCSE Food Technology NEA Vulnerable to AI?

The research and planning sections of GCSE Food Technology NEA are particularly susceptible to AI use, for a few interconnected reasons. First, the tasks are written against publicly available briefs. AQA and WJEC release their NEA tasks in advance, and students research the topic over an extended period before submitting their written analysis. AI tools can be pointed at the same brief and will produce well-structured, contextually relevant responses that cover nutritional analysis, target audience considerations, and development justifications.

Second, Food Technology written work rewards a specific type of articulate, structured argument - the kind that AI models produce fluently. A Year 10 student who struggles to explain the Maillard reaction in their own words can instruct an AI to explain it clearly, in appropriate technical language, in under thirty seconds. The result is a response that may be indistinguishable from a capable student's genuine work when read quickly across a class set.

Third, the written sections of the NEA sit alongside practical work and photographic evidence. The disconnect between a student's polished written analysis and their practical performance - or between their in-lesson contributions and the sophistication of their submitted writing - is often the first thing that alerts a teacher that something is not right. AI detection tools give you a way to investigate that concern systematically rather than acting on intuition alone.

How Likelihood Scores Work in Practice

AI detection tools analyse statistical patterns in text - the predictability of word sequences, syntactic regularity, vocabulary distribution, and structural consistency - and compare them against characteristics common in AI-generated writing. The output is a likelihood score: a number from 0 to 100% indicating how closely the text resembles AI-generated content.

A high score is not a verdict. It is an indicator that warrants further investigation. False positives do occur: students who have received very structured writing scaffolds, who have closely followed modelling examples, or who are exceptionally precise and formal writers can produce text that scores higher than their actual abilities might suggest. Equally, a student who uses AI to generate a first draft and then rewrites it substantially may produce text that scores lower than you might expect. No detection tool is infallible, and the score should always be one piece of evidence among several - not a conclusion in itself.

GradeOrbit's AI detection, part of its school accounts, takes a different approach: it compares the submission with that student's own earlier work on file and returns a score for how far the new piece departs from it, alongside the signals that contributed to it and a plain-English reasoning paragraph. That transparency is important: it gives you something concrete to discuss with a student or a senior colleague, rather than an opaque number with no explanation behind it.

Signals Beyond the Score - Contextual Judgment

When a detection check returns a high likelihood score, the most useful response is to hold it alongside everything else you know about the student and the submission. There are several contextual questions worth asking before reaching any conclusions.

• Does the written analysis reflect the same level of knowledge the student demonstrates in practical lessons and verbal discussions?
• Is the technical vocabulary and sentence structure noticeably above what the student produces in timed, in-class writing tasks?
• Are there drafts, research notes, or annotated sources that document the development of the written analysis?
• Does the student's work show specific, personal engagement with their chosen brief - or does it read as generically applicable to any Food Technology task?
• Has the student been able to discuss their analysis in conversation, or do they struggle to explain ideas they have apparently written at length?

In many cases, contextual evidence will either confirm or significantly reduce your concern. A high score on work from a student who consistently engages thoughtfully in class, produces annotated drafts, and can articulate their analytical choices verbally should be treated very differently from a high score on a neatly typed submission from a student with no draft evidence and limited in-class participation.

For more detail on navigating this process, the guide on how to interpret AI detection scores covers the decision-making framework teachers use when a score is high.

Using GradeOrbit's Detection Tool for Food Technology

On a GradeOrbit school account, run detection from a pupil's profile for one piece, or on a whole class at once. Upload scans, photos or PDFs, or use the phone QR camera link; for a class set, bulk upload splits the pages by "pages per student" and matches the name on each cover page to your class list, asking you to pick the pupil if the match isn't clear. For Food Technology NEA sections that students have completed digitally, saving them as PDFs is the most straightforward route (Word documents aren't accepted, so export to PDF first).

Checks are unlimited and included in every school teacher seat, so there is no separate deeper-analysis tier to save for the submissions you are more concerned about - the same check gives you an initial scan of a class set and, if a piece merits a formal conversation with a student, a head of year, or the exams officer, a detailed signal breakdown to go with it. The check compares the writing against that student's own earlier work already on file, which is a stronger basis for a formal conversation than a generic score alone. A pupil with no earlier work on file is skipped, and you are told. Pupils' work is held under the school's signed Data Processing Agreement for the school's retention window and is never used to train AI models.

Save the output - score, signal breakdown, and your professional notes - before taking any action. Documentation matters, both for the integrity of any formal process and for your own protection if a concern is later disputed.

Involving Students and Maintaining Fairness

If a detection check raises a concern serious enough to pursue, the appropriate first step is a professional conversation - not a formal referral. Ask the student to talk you through their written analysis: where their ideas came from, how they developed their argument, and what they found difficult. Ask them to complete a short related writing task in a supervised setting. In the large majority of cases, this conversation will give you the clarity you need without escalating the situation unnecessarily.

AQA and WJEC both have clear guidance on suspected malpractice in NEA components, and your school's academic integrity policy should set out the formal steps. Your exams officer will be familiar with the JCQ procedures. The principle throughout is that a likelihood score alone is never sufficient grounds for formal action - it is the starting point for a process that gives the student a proper opportunity to respond.

It is also worth being proactive at the start of NEA season. A clear explanation, delivered to students before the research and planning phase begins, of what constitutes AI misuse under examination regulations - and what the consequences are - reduces the likelihood of dishonest submissions. Many students who use AI tools in coursework do so without fully understanding the rules. Early, plain-language communication protects them as well as the integrity of the assessment.

Start Detecting AI in Your Food Technology Marking

AI detection is becoming a standard part of the professional toolkit for teachers managing NEA components. Used carefully - alongside contextual knowledge of your students, documentation of the drafting process, and a fair, structured response to concerns - it helps protect the integrity of qualifications that students have worked genuinely hard to achieve.

GradeOrbit's AI Detection tool gives you a transparent, documented result you can act on professionally, with the signal detail needed to inform a proportionate response.

Start a free school trial and run your first detection check on GCSE Food Technology coursework.