GPT-generated tutor feedback is predominantly positive or neutral in sentiment¶

Claim · All claims
Evidence · 1 study · design r3 · q2
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person · recheck by 2028: tool and model accuracy changes quickly

Subclaims¶

q2 i? Automated sentiment analysis found most GPT-generated hints positive or neutral, with only 292 hints categorized as negative. → Reddig 2025

Evidence¶

Reddig 2025¶

Reddig, J. M., Arora, A., & MacLellan, C. J. (2025). Generating In-Context, Personalized Feedback for Intelligent Tutors with Large Language Models. Int J Artif Intell Educ. https://doi.org/10.1007/s40593-025-00505-6

q2 · i? · design · r3

TextBlob lexicon-based sentiment analysis of the GPT-generated hints; polarity ranged from -0.75 to 1.00 with a high concentration just below 0.00, and the histogram was skewed right, indicating most feedback was neutral to positive.

"The bar graph in Fig. 8 agrees, with most feedback falling into the positive category, and only 292 hints categorized as negative."

Discussion¶

Related Claims¶

• Counterfactual fine-tuning reduces sentiment bias in LLM-generated forum replies — related
• GPT4-coded sentiment of students' open-ended general feedback about the intervention was overwhelmingly positive — related
• LLM-generated forum replies are more neutral and less varied in sentiment than human replies — related
• 90% of open-ended survey responses about the ChatGPT activity were neutral in sentiment, with 2% positive and 8% negative — related