Combine AI-based feedback on early drafts with human feedback on later drafts, and teach students to evaluate AI feedback critically¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The authors recommend deploying generative AI feedback in writing instruction in ways that maximize its strengths (instantaneous, iterative, no training corpus) while minimizing its limitations (inaccuracy, poor prioritization for high-scoring essays). They suggest instructors "can combine AI-based feedback on earlier drafts with human feedback on a later or final draft," remind students that not all feedback is accurate, and teach students to "evaluate the feedback for accuracy and usefulness before integrating it into subsequent drafts." This rests on AI literacy: teachers and students need to understand how AI functions and use it critically and reflectively.

Design Implications¶

Context¶

Requirements¶

• Teachers and students must be taught how to use and effectively prompt the generative AI
• Students must be taught that AI feedback is not infallible and should be selectively taken up

Constraints¶

• AI may not always prioritize essential features, especially for higher-scoring essays
• AI accuracy declined for high-quality essays and supportive tone was lowest for low-quality essays in this study

Target Learners¶

• secondary students (Grades 6-12)
• English learners

Target Learning Goals¶

• improving writing through formative feedback
• developing AI literacy and critical evaluation of feedback

Related Strategies¶

• Integrate ChatGPT as a feedback resource alongside peer feedback for essay writing in higher education

Examples¶

-

Key Sources¶

• Steiss, Jacob; Tate, Tamara; Graham, Steve; Cruz, Jazmin; Hebert, Michael; Wang, Jiali; Moon, Youngsun; Tseng, Waverly; Warschauer, Mark. (2023). Comparing the Quality of Human and ChatGPT Feedback on Students' Writing. https://eric.ed.gov/?q=Comparing+the+Quality+of+Human+and+ChatGPT+Feedback+on+Students+Writing