Case texas-am-commerce-chatgpt-false-accusationAI misfireUnited States
Texas A&M-Commerce instructor asks ChatGPT if it wrote essays, then withholds grades of flagged seniors
An instructor pasted students' essays into ChatGPT, asked whether it had written them, and gave flagged students a zero or incomplete grade, putting some graduating seniors' diplomas on hold. ChatGPT cannot reliably detect AI text; the university said several students were cleared, one admitted using ChatGPT, and no one failed or was barred from graduating.
MediumHarm caused by AIStatus: ResolvedVerified case file
Open in the interactive mapCite this case
- Reported
- May 18, 2023
- Location
- United States
- Target
- Students in an animal science class at Texas A&M University-Commerce
- Severity
- Medium
- Verification
- Confirmed by victim, vendor or authority
- Triggered by
- ChatGPT used as an AI-text detector by an instructor
- Actor type
- AI product design / safeguards
- Sector
- Education & research
- Reported loss
- None reported
- CVE
- None
What happened
An instructor pasted students' essays into ChatGPT, asked whether it had written them, and gave flagged students a zero or incomplete grade, putting some graduating seniors' diplomas on hold. ChatGPT cannot reliably detect AI text; the university said several students were cleared, one admitted using ChatGPT, and no one failed or was barred from graduating.
Impact: Graduating seniors' grades and diplomas temporarily withheld
Who triggered it
Actor: ChatGPT used as an AI-text detector by an instructor
Attribution: Texas A&M University-Commerce statement; Washington Post reporting
Motive: None (misuse of AI as a detector)
How it happened
Attack vector: Chatbot asked to judge authorship of student work
Root cause: Reliance on a chatbot's unreliable claim of authorship as evidence of cheating
Timeline
- Accusations and diploma holds reported; university issues statement
Response
University investigated case by case, cleared several students and began developing AI policies
Current status: Resolved
Sources (open in a new tab)
- The Washington Post via Anchorage Daily News adn.com
- Interesting Engineering interestingengineering.com
Cite this case
Texas A&M-Commerce instructor asks ChatGPT if it wrote essays, then withholds grades of flagged seniors. AI Threat Watch, case texas-am-commerce-chatgpt-false-accusation, May 18, 2023. https://aithreat.io/cases/texas-am-commerce-chatgpt-false-accusation/
Permanent link: https://aithreat.io/cases/texas-am-commerce-chatgpt-false-accusation/
Related cases
The same type of incident, most recent first.