Small edits to cGPT instructions and knowledge bases had unpredictable effects on GPT responses during iterative co-design¶

Claim · All claims
Evidence · 1 study · qualitative r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? During iterative cGPT development, students found that edits in wording, uploaded documents, and sentence additions or deletions could either clearly affect GPT responses or have no apparent effect. → Lo S 2026

Evidence¶

Lo S 2026¶

Lo S, Lea T, Mavromoustakos A, Joseph B, Yii M, Exintaris B, Karunaratne N, Pearson DL, Ritchie T, Yuriev E and Pyun J (2026) Co-creating custom GPTs: an autoethnographic study of undergraduate students as partners in generative AI innovation. Front. Educ. 11:1818781. https://doi.org/10.3389/feduc.2026.1818781

q2 · i? · qualitative · r2

Inductive thematic analysis of summer-scholarship students' SPAR autoethnographic logs written during iterative cGPT construction. Students reported that "Edits in wording, uploaded documents, addition and deletion of sentences and paragraphs" produced inconsistent effects on model responses, requiring sustained critical evaluation of outputs.

"Edits in wording, uploaded documents, addition and deletion of sentences and paragraphs could either clearly affect how the GPT responded, or have no apparent effect"

Discussion¶

Related Claims¶

• Students reported that continued interaction with a cGPT improved its adherence to instructions — related
• Students' evaluative judgment of cGPT responses shifted from strict principle compliance toward contextual appropriateness — related
• GPT responses reliably incorporated broad ideas from instructions but resisted precise control — related