Abstract
Digital education tools have revolutionized how learning solutions can be personalized for the benefit of all learners. When advances in technology are paired with knowledge of how individuals learn most effectively, more efficacious learning solutions can be developed. In this chapter, the authors contend that not only should learning science principles drive the technologies learners engage with, but combining these principles with the algorithms that personalize a learner's experience will greatly impact learning outcomes at scale. They offer a brief discussion of automation and self-regulation in educational technology, then provide an example of a novel learning solution that pairs personalization—through automated intelligence—with learning science principles to impact outcomes. A practical discussion of how to design and develop learning solutions for maximum impact is shared, as well as best practices for conducting validity and efficacy research to measure the extent to which intelligent self-regulation features are supporting learning outcomes.
TopIntroduction
Digital education has transformed how we can personalize learning at scale; and recent advances in machine learning and natural language processing have enhanced learning efficiencies, created flexibilities, and enabled dynamic and actionable insights. In some cases, however, the technology is out in front of, or developed in absence of the learning theory it is supposed to be representing–such as self-regulated learning (SRL). In this chapter, we contend that not only should learning science principles drive the technologies learners engage with but combining learning science principles with the algorithms that personalize a learner’s experience will greatly impact learning outcomes at scale.
Understanding when, and in what context certain features will support individual student success can help target these kinds of interventions and create more impactful learning experiences. Particularly when including SRL features in a tool, personalization and targeted interventions are critical. Some SRL features, like goal setting, are likely to benefit all learners and should be a standard feature in that tool and made available to all learners. Other features, like those that encourage and support self-monitoring, should be personalized. Sensitive, complex algorithms can drive how often learners are asked to self-monitor, determine whether they have the agency to engage in monitoring, and the extent to which they can monitor their learning. Models that become increasingly intelligent as a learner engages with a tool have the potential to personalize both the learning content they interact with, and the SRL features themselves.
In this chapter, we discuss the importance of leveraging learning science principles in conjunction with developing sophisticated algorithms to create meaningful SRL features in personalized digital tools that target a learner’s needs. Following brief discussions of both AI and SRL in EdTech, this chapter will be divided into three sections that relate to bringing AI and learning science together to enable automated self-regulation supports in digital solutions.
The first section of the chapter focuses on product. Using a novel learning solution that is currently under development as an example, we will provide recommendations for how learning- science-based AI can offer intelligent SRL. The product offers SRL-promoting features to learners when they are most in need of them while not overburdening learners who may not be benefiting from them. We will outline the specific features built into the example tool, share insight into the logic that personalizes the features of the tool, and discuss the impact that those features are expected to have on learning.
The example product is a classroom-based automated writing evaluation tool that is designed to be used in middle school classrooms. It helps learners plan their writing, connect that planning to their writing, and create multiple drafts based on automated evaluation and feedback. This tool was selected because engaging in, progressing through, and completing the writing process can be a very challenging task for learners. Research suggests that learners tend to become frustrated during the writing process and potentially disengage if they lack confidence or feel that they are not successful in the instructional strategy.
The second section of the chapter focuses on product development. We will discuss considerations for the research and development of SRL tools, focusing specifically on the importance of successful collaboration amongst the teams responsible for designing and building features into the products. We posit that a Scrum team that includes Learning Scientists, User Experience Researchers, Software Developers, and AI/NLP Engineers who are sprinting together will build the most effective tools. Our contention is that fully dedicated cross-functional teams making product development decisions together will ensure that meaningful, automated SRL features are prioritized. Continuing with the product examples shared in the first section of the chapter, we will discuss the successes and challenges of those teams and make practical recommendations for Scrum teams working cross-functionally.
Key Terms in this Chapter
Student-Led Product: A digital learning tool that students are meant to be able to successfully engage with and complete without direction from the teacher.
Intelligent Self-Regulation: Surfacing features to a learner within a digital solution that support their ability to control their own learning process.
Learner Motivation: A learner feeling encouraged to engage in and persist through an instructional strategy.
Learner Confidence: A learner believing that they can successfully engage in and complete an instructional strategy.
Automated Writing Evaluation: The use of natural language processing to automate the scoring and feedback of written text.
Natural Language Processing and Scoring: The use of textual or spoken language to model learner performance and provide signals about that performance.
Learner Insights: Applying an interpretation to signals provided by scoring to help learners contextualize the scores they are provided and action the insights.