AI Tools in Education: Learning, Cognition, and Ethics

AI Tools for Tracking and Monitoring Learners

There are a wide range of AI tools that have been deployed in attempts to not only track but also monitor learners. Thanks to these monitoring capabilities, "AI can customize the 'feed' of information and materials into the course according to learner's needs, provide feedback and encouragement" (Popenici and Kerr, 2017, p. 10). AI has thus created a sound platform for the further advancement of individualized and differentiated learning. Thanks to AI, educators can now assess the unique needs of learners and make relevant adjustments to the learning process—something that would be difficult to achieve without it. Among the enterprises that have made significant progress in developing meaningful AI platforms are Carnegie Learning and Content Technologies (Panconesi and Guida, 2021). According to the authors, these companies offer "digital platforms that use AI to provide learning, testing and feedback to students from pre-K to college level that gives them the challenges they are ready for, identifies gaps in knowledge and redirects to new topics when appropriate" (Panconesi and Guida, 2021, p. 311).

As Farhadi (2018) points out, AI has also made it possible for relevant pedagogical methodologies to be developed in order to closely monitor student learning. Some of the AI techniques and approaches of particular relevance on this front revolve around "automated planning and scheduling, and how they can be applied to pedagogical and educational environments" (Farhadi, 2018, p. 71). One tool that proves useful in scheduling efforts—and therefore in tracking the learning process—is AI-powered school scheduling software. In essence, such tools make it easier for schools to prioritize crucial student needs while simultaneously helping to "visualize shifts to one's schedule, and preview the impact on teacher and student experiences" (ABL, 2021). These tools are instrumental for academic institutions seeking more effective scheduling solutions, as they help expand staff capacity, resolve challenging scheduling conflicts, and ensure that district goals are incorporated appropriately.

Any assessment of AI tools for tracking and monitoring learners would be incomplete without considering the active role of machine learning in this domain. Machine learning has made it possible for instructors to monitor learner progress via tools such as Panorama Education (Panconesi and Guida, 2021). This particular tool enables educators to make use of specific student data for decision-making purposes. More specifically, "educators use the platform to monitor how pupils are faring in academics, attendance, behavior, and college readiness and then coordinate actions to support each child" (Panconesi and Guida, 2021, p. 347). Machine learning concerns the various approaches to data analysis through the automation of analytical model building. A tool such as Panorama Education actively seeks diverse insights from the rich data sets it encounters and identifies patterns that may be deemed meaningful. As a result, it is able to provide instructors with visual dashboard reporting by pooling together relevant student data (Panorama Education, 2021). Instructors can then use this information to take necessary action aimed at improving learner outcomes. Those who have worked with the tool note that it significantly reduces the time required to track certain student progress metrics. For instance, one school district assistant principal observed: "It used to take me two to three days to pull the data across attendance, behavior, and grades with spreadsheets. It now takes minutes using the Panorama system" (Panorama Education, 2021).

AI Tools for Detecting Cognitive Issues

In a cognitive learning environment powered by AI, the tools deployed are designed to determine a student's preferred learning approaches, the student's level of interest in a given topic or area of study, and the extent to which the student is competent in a particular subject area (Zawacki-Richter, Marin, Bond, and Gouverneur, 2019). It is on the basis of this information that an ideal, personalized learning approach is formulated for each learner. With continued improvements in AI, we are likely to transition away from traditional approaches to cognitive learning, which have been somewhat limited.

According to Zawacki-Richter, Marin, Bond, and Gouverneur (2019), cognitive learning has historically depended heavily on repetition, in which students are trained continuously with training intensity kept constant or gradually varied over time. Although this approach has been associated with improved performance, not all participants are able to achieve optimal outcomes. Consequently, other strategies for promoting cognition have been explored, including mental exercises, video games, and drug treatments (Zawacki-Richter, Marin, Bond, and Gouverneur, 2019). In recent times, and thanks to significant advances in technology, digital therapeutics have also emerged as a promising avenue. It is important to recognize that all individuals are unique, which accounts for performance variances and limits the effectiveness of conventional learning strategies. AI tools for detecting cognitive issues address this gap directly. A notable example is CURATE.AI, which has been shown to be "particularly instrumental in efforts to customize training regimens for individuals to personalize learning and improve cognitive performance" (Habib, 2019, p. 211). The tool makes active use of a specific learner's performance data to build an individualized profile, enabling the adaptation of cognitive learning to that learner's specific competencies and habits. The data used for calibration includes performance scores on relevant tasks and measures of training intensity (Habib, 2019). However, as noted elsewhere in this review, some researchers have criticized AI systems that seek to adapt the cognition of students—especially when those systems employ what Schiff (2020) refers to as "potentially addictive gamification techniques" (p. 342).