ID Tip: Using AI for Assessment Can Make More Time for Human Connection

This week’s tip is inspired by Professor Zhongzhou Chen’s recent talk, Leveraging Machine Intelligence for More Human Touch in Teaching and Learning, hosted by the UB Office of Curriculum, Assessment, and Teaching Transformation.

Professor Chen’s research shows that when used thoughtfully, AI can help faculty spend less time grading and more time engaging with students. By integrating assessments and rubrics directly in Blackboard Ultra, you can automate portions of feedback and grading while maintaining control over quality and alignment. The result is more time for what matters most, mentoring, answering questions, and building meaningful connections.

Best Practices for Leveraging AI in Assessment

1. Leverage Built-In Tools
   Use your LMS (e.g., Blackboard Ultra) to build assessments with integrated rubrics and automated grading or feedback options. This saves time on routine/administrative tasks and allows you to focus on higher-value interactions.
2. Stay Transparent
   Clearly communicate to students how their work is being assessed, including any automation or AI-assisted processes. This transparency builds trust, clarifies expectations, and helps them engage with the criteria and feedback meaningfully.
3. Focus Where It Matters
   Identify which tasks benefit most from your direct intervention, such as conceptual feedback, scaffolding student reasoning, or supporting metacognition, and apply automation to more routine or structured parts of the assessment workflow.
4. Align Practice and Assessment
   Chen’s research emphasizes the importance of alignment between practice tasks, feedback processes, and assessment items. When automated tools are used, ensure that the scaffolded practice, the rubric, and the grade/feedback workflow all map to the same learning goals.
5. Use AI to Augment Instructor Expertise
   As seen in the studies of GenAI-assisted feedback and grading, AI can generate draft feedback or evaluate responses, but many tasks still benefit from human review or refinement. For example, one study found that AI-generated feedback sometimes was rated as “more human-like,” but needed minor instructor edits to ensure accuracy and appropriateness (Chen et al., 2024).
6. Monitor, Validate, and Iterate
   When deploying automated or AI-assisted assessment workflows, routinely monitor outcomes (grading consistency, student perceptions, learning gains). Adjust prompts, rubric alignment, and instructor involvement as needed. Chen’s team found that model accuracy depends on clarity of criteria, iteration, and instructor oversight (Chen & Liu, 2023).
7. Preserve the Human Touch
   Ultimately, the goal is not to replace instructor-student interaction but to free up time for it. Use the time saved by automation to engage students in meaningful dialogue: hold thoughtful office hours, respond to conceptual questions, facilitate peer-discussion, and build relationships that support motivation and belonging.

Learn More

Chen, Z. (2024, April). Leveraging machine intelligence for more human touch in teaching and learning. University at Buffalo Office of Curriculum, Assessment, and Teaching Transformation. https://www.buffalo.edu/catt/teach/this-semester/excellence-in-teaching.html

Chen, Z., & Liu, C. (2023). Human-AI collaboration in assessment: Amplifying instructor feedback and student learning. Physical Review Physics Education Research, 19(4), 010152. https://doi.org/10.1103/PhysRevPhysEducRes.19.010152