AI Use in Assessments

AI is changing how teachers and professors give feedback to students and how their work is graded, Alongside this, AI is changing how students complete assignments and implications for their future careers and real world use.

AI for Feedback

The biggest advantage of using AI is the speed it offers. A student can submit a writing assignment and get targeted and specific feedback in seconds rather than having to wait days or weeks for educators to grade and respond to their work. Adaptive AI softwares can also help students learn by increasing difficulty based on their performance.

Though AI feedback can be a very useful tool, it has its drawbacks. It can general do surface-level analysis well like grammar or compare writing to a rubric, but it can struggle with nuance. It cannot account for creativity, intellectual thinking behind work, and how people's work has increased in quality over time. Another large issue with AI is its hallucinations, which is when it confidently says statements that just are not true.

Academic Integrity and Ethics

AI tools only know as much as the data they are trained on. Inside of these data training sets (a large selection of the internet) there are inherent biases across race an socioeconomic lines. This issue becomes even greater when trying to "automatically score an essay" due to these innate biases in the program. AI plagiarism detectors have produced false positives which can hurt students who completed their assignment with integrity. 

The line between "AI-generated" and "AI-assisted" is becoming blurry and as a results students are completing assignments within this grey area. Institutions need to have evolving policies that reflect how modern AI systems can be implemented within assignments

Human judgement should remain central for final grades and important assessments where small inconsistencies may matter a lot (think SAT or ACT).

My Perspective

If it were up to me I would use AI as a general check for the obvious things like grammar, spelling, and rubric alignment for less important assignments, like first drafts of papers. This allows students to get more timely and in depth feedback. For graded work, AI can assist by being a tool that can flag areas of interest for professors and teachers to look more into.

Boundaries that should be set are never allowing full autonomous grading of assignments, just use it for feedback and flags