CASE STUDYUniversity of North TexasImproving Writing Quality and Reducing Grading Workload Through Human–AI Collaboration at The University of North TexasA peer-reviewed study from the University of North Texas, published in the Journal of Computing in Higher Education (2025), explored how Packback Writing supports instructors and students through AI-enabled feedback loops. Across four courses and more than 2,800 long-form assignments, researchers found that Packback improved the mechanical quality of student writing while significantly reducing the time faculty spent grading. The study demonstrates how thoughtful human–AI collaboration can enhance feedback, equity, and writing mastery at scale. Key Takeaways Writing Mechanics & Structure: Students’ scores improved by +6 % in Flow & Structure, +3 % in Grammar & Mechanics, and +3 % in Depth/Word-Count over the semester. Grading Alignment: There was a strong correlation (≈ 70 %) between AI-suggested and instructor-assigned grades, with instructors adjusting scores by only ~2–5% in most cases Time Savings for Instructors: Faculty reported up to a 50 % reduction in grading time, enabling them to invest more in personalized, higher-order feedback and mentoring. Equity & Consistency: The automated baseline feedback helped a diverse student cohort establish stronger writing foundations, allowing instructors to elevate class wide writing quality without uneven workloads. Situation At the University of North Texas, instructors of four courses (psychology, public administration, and learning technologies) faced persistent challenges: Students submitted long-form writing assignments that often struggled with mechanics (grammar, structure, depth) rather than only with ideas Faculty spent substantial time grading and providing feedback on low-level errors, leaving less time for higher-order feedback on argumentation, clarity, and critical thinking The institution sought a scalable way to enhance writing outcomes for a diverse student population while preserving instructor time and ensuring consistency in feedback Action Packback Writing (via Packback Deep Dives) was implemented across those courses under a human–AI collaboration model Students used the platform during drafting, receiving real-time AI-powered feedback on structure, grammar, citations, flow, presentation, and depth Instructors continued to oversee the writing process, focusing on the “Content & Ideas” dimension while the AI handled five other rubric dimensions (Flow & Structure; Grammar & Mechanics; Formatting & Presentation; Research & Citations; Word Count & Depth) Faculty were trained to interpret AI suggestions, override them when necessary, and spend saved time on more sophisticated feedback Over 2,800 assignments from 449 students were graded both by the AI and by instructors, enabling quantitative alignment analysis and qualitative interviews to assess perceptions of workload reduction and writing improvement Result Results from the multi-course pilot demonstrate how Packback’s human–AI collaboration model improved writing quality and instructional efficiency including: Writing proficiency rises + instructor bandwidth expands: Improved student writing and reduced grading burden occurred together, not at the expense of either Scalable quality across diverse students: Automated feedback ensures consistent foundational support, helping instructors scale high-quality writing instruction across classes and student populations Implementation matters: Successful outcomes depend on integrating the tool into the drafting process (not just post-submission), training faculty, and maintaining instructor oversight of higher-order skills Human–AI partnership works: When AI handles mechanical feedback, instructors can focus on what only humans can do like critique ideas, inspire depth, and motivate growth Real Impact, Real Voices The AI’s presence sped up my grading twofold. I spent less overall time worrying about grammar and more time focusing on content. Faculty Participant University of North Texas SCHEDULE A DEMOWant to acheive similar results ? Contact us today. "*" indicates required fields First Name*Last Name*Email* Your Role*Select OneFaculty (Higher Education)Administrator (Higher Education)Instructor (K-12)Administrator (K-12)OtherInstitution*EmailThis field is for validation purposes and should be left unchanged.
