Detect machine-generated writing by checking citations, originality, factual errors, grammar, contextual awareness, and with language-analysis tools¶

Strategy · All strategies
Evidence · 5 claims (4 for, 1 against) · 4 studies (2 review, 2 theoretical), q1–q3 · 0 of 4 report an effect size · 5 claims rest on one study

Description¶

The article enumerates signs that work was written by a chatbot: repetitive or odd language patterns; missing or improper citations, since chatbots "are not capable of conducting original research or producing new ideas"; similarity to existing sources; factual errors or inconsistencies; unnaturally error-free grammar and spelling; and generic, less contextually aware text. It also recommends language analysis tools such as the GPT-2 Output Detector Demo.

Design Implications¶

Context¶

Requirements¶

• Manual review or plagiarism detection tools, plus in some cases natural language processing or machine learning algorithms trained to recognise AI-generated content

Constraints¶

• Grammar and spelling cues vary depending on the quality of the AI language model and its training data; the article cautions detection will be an arms race

Target Learners¶

• higher education students

Target Learning Goals¶

• academic integrity
• accurate assessment of student understanding

Claims¶

• Most of the article's body text was generated by ChatGPT from author prompts, and all references ChatGPT inserted were fictional [+W]
• OpenAI's GPT-2 Output Detector distinguished real student essays from ChatGPT-generated essays on the same subjects in a December 2022 trial [+M]
• Independent evaluation found commercial AI-text detectors neither sufficiently accurate nor sufficiently reliable for high-stakes use [-M]
• Large language models can produce fluent but factually inaccurate or fabricated outputs (AI hallucination), posing risks for uncritical users in science education [+M]
• AI output errors are frequent: 44.5% fabricated references in drafted research proposals and 5–13% unsafe chatbot medical answers [+M]

Related Strategies¶

• Prevent AI-assisted plagiarism through student education, signed declarations, draft submissions, clear guidelines, and close monitoring
• Redesign assessments to require critical thinking, originality, interactive activities, and real-time or proctored formats

Examples¶

-

Key Sources¶

• Debby R. E. Cotton, Peter A. Cotton & J. Reuben Shipway (2024) Chatting and cheating: Ensuring academic integrity in the era of ChatGPT, Innovations in Education and Teaching International, 61:2, 228-239. https://doi.org/10.1080/14703297.2023.2190148