Fulfilling the promise of personalized education
director of education solutions for K12, Fuel Education
For decades, widespread personalized learning has been the dream of teachers, administrators and students alike. Personalized learning is a research-tested way to provide top-quality instruction and ensure every student fulfills their potential, regardless of their initial proficiency. Until now, however, delivering that education to every student has been difficult, or even impossible, due to the time-intensive nature of one-on-one tutoring. But thanks to artificial intelligence, personalized education for all may now be possible.
The benefits of personalized education have been proven again and again. Benjamin Bloom famously found in 1984 that individually tutored students would achieve better outcomes than 98 percent of their peers. More recently, a 2015 RAND Corporation study showed students at schools with personalized learning policies showed percentile gains of 11 percent for math and 8 percent for reading compared to their peers. And schools that had practiced elements of personalized learning for three years or more showed even greater results.
So-called “adaptive learning”—the use of algorithms to assess students’ level of mastery and areas where they need additional instruction—promised to take personalized learning to the next level through automation. But adaptive learning has so far been limited in complexity and in scope. Many tools do little more than put students on predetermined paths based on a simple assessment. Recent tools are more advanced, but a 2016 EdSurge report showed that no tool provided full adaptivity for all three stages of personalized learning: content, sequence and assessment.
But new, complex AI technologies promise to revolutionize the field of personalized and adaptive learning, finally providing an individualized education that responds to every student’s needs in all three stages. These technologies can track student progress and continuously adjust students’ content according to their particular needs. AI will support teacher planning with real-time assessments, and allow them to focus more time and energy on the parts of teaching that are most rewarding.
One of AI’s greatest contributions is in the field of content. When a student appears to be struggling at any point, an adaptive learning system can provide individualized feedback, such as a hint to why the student picked the wrong choice, instead of a binary response of “correct or incorrect.” When students fall behind, adaptive learning platforms can identify their common distractors and either provide support to bring them back on-level or target them for one-on-one teacher interventions.
The data show how effective this strategy is: a new study from Johns Hopkins, released in January, shows that completing a single AI-determined objective can lead to a 7 percent increase in the average score for content related to that objective.
AI is also useful in creating or adjusting learning sequences for students. One of the problems with the conventional classroom model is the issue of sequence—teachers must teach the same content to all students in the same order, regardless of their knowledge level. But AI can adapt sequences to individual needs, focusing on student mastery in every core competency. That means a student in need of additional support can work on core skills, while other students are free to move ahead and progress. In addition, teachers can use the information on students’ progress to create class plans, group projects or individual instruction that target their students’ demonstrated areas of need.
After providing content and determining sequence, AI can continuously and accurately assess student performance on assignments, determining proficiency, completion rate, skill levels and more. These systems create a “backpack” of information that follows the student from class to class and from year to year. Instead of restarting the learning and assessment process every fall, students will return to school with a wealth of information to help teachers instruct them on a personal level.
AI is the key to solving a problem identified decades ago. We know that children learn best with one-on-one, personalized instruction focused on mastery. But classrooms have faced obstacles to implementing that instruction — until now. Artificial intelligence can create a unique education, tailored to student’s individual needs, while expanding the capacity of teachers, allowing them to spend less time on grading and assessing and more time on individual instruction and assistance.