The Impact of AI on Learning Analytics: Opportunities and Challenges
The use of Learning Analytics has given educators a chance to study the data resulting from students’ interactions with digital learning resources to get insights into the learning process. This field is undergoing a revolution with the availability of generative AI tools such as ChatGPT and offers novel opportunities and challenges.
Classically, learning analytics has been about collecting data from online actions such as clicks, discussions, and tests. The primary use is to enable teachers to locate students who could be described as ‘bored’ or maybe students who are having a hard time with particular lessons. However, most teachers are struggling to work with the multiple and multiple-layered dashboards, and, therefore, to decipher the data. They have also noted that new AI tools are likely to assist in closing this gap because the new tools work as translators who present the data in simplistic ways. For instance, one can imagine that AI-enabled chatbots can better explain analytics to teachers so that they can use the data without needing to be experts in analytics.
Perhaps the most promising use of generative AI in learning analytics is the capacity for the system to analyze and classify students’ contributions in the discussion forums. Earlier, this process involved a lot of work to determine whether a post was on-topic or off-topic. This task is one that can now be done fast and without much input and with a lot of accuracy when using AI. With large language models such as ChatGPT, a great deal of student work can be processed and translated into usable information for teachers.
They are also being used to improve the delivery of personalized learning. It makes sense to track a student’s activity and achievements to be able to adapt a tutoring system to the student better. Learning analytics can then feed data into these types of systems to enhance the effectiveness of personalized tutoring. This makes it easy for the system to address individual requirements of the students that is very vital as the world shifts to more personalized learning systems.
In assessments, generative AI has a great opportunity to replace traditional multiple-choice questions with more open-ended, creative questions. Computerized grading of essays or fill-in-the-blank questions has been found to be as accurate as human grading. This might even help to reduce the thinking that is usually done in the boxes of ‘yes’ and ‘no’.
However, the use of AI in learning analytics presents two issues that are bias and transparency. Like any other AI systems, generative models such as ChatGPT embody the bias, depending on the information fed to the model. It is believed that algorithmic decisions could be quite similar to discrimination and cultural or racial bias might be upheld when algorithms are employed in crucial educational settings. Thus, even when still fascinating, there are some issues with the way the conclusions are made since some of the AI systems are not clear enough to explain how this is done; this creates even more problems of trust and fairness.
AI tools also raise crucial concerns about power relations in education as these come up. More so, as AI is involved in decision making, there are questions on who owns the algorithms that shape student learning. This could make agency with teachers and students and place it in the hands of the authors of the AI systems.
All in all, with the help of AI, learning analytics is progressing as the utilization of learning assets is being optimized as well as the individual educational process is being adjusted. However, its integration must be controlled to avoid the following tools being used in a biased manner or in a way that will reduce the public’s confidence in education systems.
Source:: edsurge.com
