GenAI is here to stay. Last year, ChatGPT dominated conversations, but almost every tool has now integrated AI in some manner or form.
The use of AI grading for student assignments and exams is gaining traction among instructors. AI grading systems have proven to provide more personalized and scalable feedback compared to traditional manual methods. Mainly because most instructors do not have the time.
Incorporating AI grading and feedback offers significant benefits for both instructors and students. For instructors, it reduces the workload and ensures consistency in grading. For students, it provides timely feedback which is crucial for their learning process (University of South Carolina).
However, there are still ongoing discussions about its accuracy and bias as noted in a recent article by the Hechinger Report: https://hechingerreport.org/proof-points-asian-american-ai-bias/ where Asian students lose more grades when AI is used for grading.
The major benefit of a faster turnaround time for assignments is speeding up the feedback loop for students.
Instructors, assisted by AI, can deliver almost immediate feedback to students, promoting a faster and more engaging learning environment. Imagine drafting up a submission and then getting feedback immediately.
Students who receive feedback earlier are more likely to absorb feedback. In one of our pilots, all student teams received outstanding grades compared to teams from a past term when little or no feedback was provided.
Early feedback also alleviates the stress of uncertainty for students who might be unaware if their work will be positively received by their instructors.
It is important to be transparent with students on the usage of AI and explain how it will potentially benefit them and what are the risks. For example, AI can hallucinate facts and it is important for students to fact-check.
Effective rubrics are essential for AI grading tools to grade accurately.
Grading rubrics should include clear grading criteria, allowing the AI to evaluate student work consistently. This ensures that the feedback provided is relevant and useful for student improvement. Well-defined rubrics enhance the accuracy of AI grading by more than 30%.
We also understand that rubrics are time-consuming to create and oftentimes require the expertise of an instructional designer, that’s why we created an AI rubric generator that creates rubrics specifically for AI grading.
AI grading tools like TimelyGrader can be integrated into various LMSs such as Canvas, D2L, and Moodle. This allows educators to seamlessly integrate AI tools into their existing workflows. This is also true for students.
Using AI tools alongside LMSs also helps in maintaining data security and student privacy. It ensures that only authorized users within the university ecosystem can access course content and sensitive user information.
To conclude, there are obvious positives to using AI in the classroom but there are also risks. To mitigate risks and potential harm to the students, it is important to keep an open mind while thinking about the ‘full circle’ student and instructor experience.
Enhancing education with AI-powered grading and feedback.
