evidoria

← Back to browse

Good practice Imported

WAEC Detects and Withholds Results for AI-Generated Answers Across 235 Ghanaian Schools

Ghana · Accra · See the Ghana profile

WAEC identified AI-generated exam answers - verbatim chatbot disclaimers copied into WASSCE 2023 scripts - at 235 schools, withholding results while offering candidates a defence process, in the exam body's first documented AI-cheating case.

Details

Promoter
West African Examinations Council (WAEC), Ghana
Period
August-December 2023
Keywords
secondary education, national examinations, academic integrity, AI-misuse detection

Description

During marking of the August 2023 West African Senior School Certificate Examination (WASSCE), examiners working for the West African Examinations Council (WAEC) in Ghana noticed candidates at 235 schools writing answers that were clearly copied verbatim from AI chatbots, including disclaimer phrases such as "I'm not familiar with the term you have used" and "I cannot detect the term you have used" - responses to typed prompts, not to the exam questions as printed. WAEC's Head of Public Affairs, John Kapi, said this was "the first time WAEC will find out about this kind of irregularity in the exam," indicating some candidates had typed exam questions into mobile phones during the exam hall.

WAEC withheld the affected candidates' results pending investigation, notified them by SMS or newspaper publication, and invited them to appear before a committee of examiners to explain the irregularity before a final decision to release or cancel results. Results for implicated candidates were subsequently withheld or cancelled. WAEC also reported monitoring websites that published leaked exam questions during the sitting, which appeared to originate from photographs taken by supervisors, invigilators, or candidates.

In response, WAEC said it would work with the Ghana Education Service to tighten enforcement of existing rules banning mobile phones from examination halls, rather than deploying new detection technology of its own.

This is a reactive, enforcement-only response with no independent verification of the false-positive rate of WAEC's pattern-matching approach (spotting AI disclaimer text) and no published data on how many candidates were ultimately cleared versus cancelled; it demonstrates a real, nationally significant integrity gap opened by generative AI rather than a validated assessment tool or pedagogical practice.

Read the full analysis: https://www.legit.ng/education/1569861-how-waec-caught-students-ai-answer-questions-ghana/

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful