Dan Brown, an American author, in his novel Origin, states that “the thing that's going to make artificial intelligence [AI] so powerful is its ability to learn, and the way AI learns is to look at human culture.” In resonance to the quote, it is observable that AI-led assessments are not only prompting a revamping of traditional modes of testing and providing deeper insights to support student learning process, but also learning from data the humans enter as prompts in order to define specific strategies and enhance the assessment and testing processes for varied skill sets.
How is AI transforming assessment practices?
Assessments in the past have heavily relied on grades and marks but AI assessments are ushering in a new era of Natural Language Processing (NLP) that doesn't just check answers for factual accuracy but also processes written responses given by the students in terms of coherence, conceptual understanding, and argumentation quality. It is a wonderful tool that can be utilized by educators to evaluate summative assessments with considerable reliability.
One of the biggest advantages of leveraging AI in assessing students is its ability to decipher patterns in student learning. AI algorithms are being innovated to the point where they can identify knowledge gaps in student learning based on large datasets of student performance. This can help educators devise customized learning plans to cater to each student’s learning needs.
In the book, The Fourth Education Revolution by Anthony Seldon, the author points out that unlike humans, AI doesn't get tired or frustrated when a student takes longer to understand a topic or concept. It is this adaptive ability of AI that renders it as a game changer in the domain of assessments. Adaptive AI testing adjusts question difficulty in real-time based on student performance, providing a more precise evaluation of skill sets.
Another important aspect that AI Assessments addresses is quick and accurate feedback to students, which is essential to remedy mistakes as they are performed. This leads to deeper and more focused learning. In the past decade, even though learning management systems (LMS) were being developed from the viewpoint of helping educators to optimize learning, grading and marking assessments were totally dependent on the educators, which meant the feedback received by the student could be delayed due to unforeseen circumstances.
Furthermore, since AI can generate feedback on assessments at a much faster rate than humans, it also implies that it can create customized learning plans for each student based on the feedback it generates, helping educators focus on struggling topics for each student. Sal Khan describes the “Two Sigma Problem” which emphasizes the benefits of one-to-one tutoring in student's performance. AI in that regard has come to realize that dream to provide an impetus to student's learning outcomes.
When it comes to assessing higher order thinking skills combined with creativity and other skills related to real world competencies, AI has surged ahead with the advancement in simulations and other virtual environments, providing authentic experiences to problem solve and make decisions for real life scenarios. AI simulated environment examples include medical trainings, professional skill development, virtual reality combined with natural language processing (NLP), engineering and manufacturing simulations, software skills training, etc.
AI assessments have also opened up the possibility of creating more robust, detailed, and personalized rubrics that can be assessed by AI for critical thinking, collaboration, and creativity on a much more measurable scale than before. Automated scoring to open ended questions and performance tasks means that the teacher can focus more on the skills that need to be administered rather than the dreadful grading that used to take a big chunk of an educator's time.
Even though some argue that generative AI is being used as a cheating tool, on the contrary, AI assessments have improved plagiarism detection via algorithms that can check for potential academic dishonesty concerns by finding patterns in student's submitted work vs existing content on the internet. AI based proctoring where a software continuously observes a user taking an exam via webcam and tries to identify if a user is trying to cheat by opening another tab or looking into a book or notes by his/her side has proved to be an excellent step in preventing cheating on online exams and tests.
Even though the use of AI has far more potential in the context of creating assessments, there are aspects that school administrators and teachers must consider while promoting the use of AI assessments.
The biggest challenge is the transparency of use of AI in making decisions on assessments in school settings. This is still not made very clear by various companies that have created generative AI software. In order to create policies, administrators and other stakeholders in educational institutions need to have transparency over how AI makes decisions for the assessments. Transparency fosters trust which doesn't exist yet, especially when it comes to sensitive issues of grading and evaluations. Another aspect to consider is the privacy of data that AI assessments may compromise in trying to create customized learning plans for each student, leading to data security concerns, etc. Lastly, because AI creates content based on data available on the internet, it is susceptible to biases while creating assessments for students.
Although, there are some factors that educators and administrators must consider while choosing to adapt AI in creating and evaluating assessments, it is no wonder that AI is here to stay and should be embraced for the potential it brings to foster more focused and individualized learning. I would like to leave you with some food for thought, what if one day you walked into a hospital and the nurse just asked you to scan your finger with the biometric scanner and the doctor could get access to all the information from all the hospitals you visited in your life and told that he knew everything about your medical history, would you consider AI a boon or a curse?
References
https://blog.khanacademy.org/sal-khans-2023-ted-talk-ai-in-the-classroom-can-transform-education/
https://www.prodigygame.com/main-en/blog/learning-gaps/
https://www.geeksforgeeks.org/pattern-recognition-introduction/
https://profiletree.com/using-ai-simulations-for-hands-on-training/
https://vrvisiongroup.com/the-intersection-of-vr-and-ai/
--------------------------------------------------------------
Uma Shankar Singh is a science, technology, engineering, arts, and math (STEAM) and Computer Science teacher at St. Johnsbury Academy in Jeju, South Korea. He likes to read about human psychology, troubleshoot computers and other electronics, create projects with 3D printers and fly drones.
LinkedIn: Uma S. Singh
