Editor’s note: This essay is an entry in Fordham’s 2023 Wonkathon, which asked contributors to answer this question: “How can we harness the power but mitigate the risks of artificial intelligence in our schools?” Learn more.
We live in a world where physical textbooks are dead like the paper they were printed on. Despite my affinity for hardbound books with signatures from older brothers and sisters written in the front cover along with a list of twenty-five other students, this type of instruction is fading fast, with its own timely end, expedited by the recent pandemic.
We’ve been integrating AI into classrooms since the first adaptive assessments became part of end-of-course exams. Today it’s even commonplace to have adaptive homework, tutoring, and in-class assessments on a “one to one” device basis. This type of AI in education relies solely on an input method focusing on the student.
Plenty of educational vendors promise results to narrow the achievement gap, which existed long before Covid. Many AI platforms have been in existence in classrooms for decades at the best schools, with the well-funded budgets, who offered “one to one” well before the feds required it for standardized testing. But our students’ achievement hasn’t kept pace with the promise of these technologies.
The answer to our AI innovation in education isn’t solely a student-centered approach with inputs only and teachers as device monitors. In fact, AI is safest and most effective when it is focused on the adult in the room, instead of minor children. As existing achievement gaps persist for the student, we should be focusing on an input model for the teacher. It has been widely written that the impact on students of having a high-quality teacher in the classroom cannot be ignored no matter what content you have on a device.
Enter robot teachers, the next generation of education, requiring no sleep, no sick days, and no student loan debt to pay off. They can harness the power of the infinite knowledge available through an internet connection and customize to every single student’s needs. Common to other types of educational software on the market, robot teachers would have real-time feedback from student assessment, content to accelerate or catch up students who are not ready for what’s next, and a dashboard for the teachers to see how students are doing.
Yet robot teachers, let’s call them AI coaches, go beyond what is part of the classroom today by helping teachers become high-quality teachers in real time, without going back to school, with the students they have in class today. Using the best pedagogy methods, human psychology, and research-based methods of instruction in real time, the teacher becomes a pilot navigating their classrooms with the benefits of AI.
Leveraging the endless information already available, AI coaches will curate content to assist the teacher in inspiring students to fully dive into what they are learning and develop a deeper level of inquisition for the student sans Google search bar. Curriculum linkages both horizontally and vertically to other subjects throughout the school, AI curricular mapping, and real time progress that provides feedback to teachers will help automate the high-quality teaching throughout the school year, reducing boring hours in irrelevant teacher professional development and associated costs.
Using Ralph Tyler’s four-phase curriculum cycle of analysis, design, implementation, and evaluation (Tyler, 1949) as a basis for the AI coach process design, a camera follows the teacher while she is instructing students doing an analysis based on speech, body language, and delivered content. Students perform AI created tasks on a device, which then provides instant feedback to the teacher, which many are accustomed to today, and informs the AI coach’s design process to prompt the teacher immediately with review, teaching strategies, or new content.
During the “analysis phase,” as AI coach records the teacher instructing, students would respond to quick assessments throughout the class. Instantaneously, given the nature of AI, the teacher would receive AI coaching. This coaching would provide a high-quality example for the teacher to work through with the students, along with a new short assessment for the students to see who now understands the material.
With the precision of a surgeon, the teacher could focus on students who didn’t understand the concept using another example tailored to their current understanding of the subject, while other students were guided to explore the topic further to read or write more on the subject and/or do additional problems, with no lost time for students who already are proficient on the topic.
Using the data from the analysis phase, the AI coach will design and recommend the next steps for the teacher in real time. Options include the teacher moving on to new material or additional content based on its horizontal and vertical integration of what is happening in other classes, subjects, or grades through a high-quality example. It can also recommend students take a break with content for the break, or provide a high-quality example for the teacher to work through with individual students while splitting the remaining students into countless groups or individual assignments tailored to each student’s progress and interests.
During the “implementation stage,” the teacher is working through a high-quality example with students who need it, while, unlike today, the rest of the class isn’t left idle with lost learning time, but instead, using this horizontal and vertical framework across grades and classrooms, the AI coach would provide additional content. For example, for a student in sixth grade math who is doing well on this topic, the AI coach would present basic Euclidean theory, a short biography of Euclid as a preview for upcoming geometry class, or a battlefield from WWII that used this math to determine where to drop resources linking it to a topic studied in social studies. For students who need additional help, the AI coach would provide high-quality coaching, including other methods of teaching or other high-quality examples with linkages to other subjects that increase comprehension and move towards proficiency.
Another benefit to teachers during this phase is AI device control that keeps students from navigating away from the content. If a student tries to navigate away, the student can be reminded of why it’s important to stay on topic, and could even suggest a break for that individual student, including standing up and stretching or a “brain break” activity. In addition, it could ask the student to self-reflect on their work in the class, other classes, or another topic entirely.
Not only will AI coaches help students in real time, during the “evaluation phase,” the teacher will reflect using collected data, experiences in the classroom, and be able to review videos created by the AI coach with suggestions on how the lesson could have been implemented better or additional strategies that the teacher could use in the future. These videos would be of the teacher, but created by the AI coach so they can see themselves giving the lesson. The teacher could use this tool for reflection whenever they wanted, daily or weekly. The teacher could include notes on which subjects need to be reviewed again based on the data and get insights on which kind of teaching methods work for each of their students. The AI coach could even design assessments based on the data and align them to standards negating the need for the use of a curve. The data has endless possibilities for the teacher and school leadership.
Principals would review a dashboard of data daily and/or weekly looking at real-time student progress and recognize teachers who are excelling more frequently. A principal could also use these data for curricular implementation, points of content intersection across the school, and to design new innovative programing for the students. Teacher evaluation would become more useful and simpler, and training a substitute teacher corps on the AI coach means that teachers can actually take days off without student learning loss.
While technology will continue to fill in the gaps through tutoring and curriculum integration, it won’t replace having a high-quality teacher at the front of the classroom, which our country desperately needs. Covid dollars will run out, and trained high-quality teachers will still be needed. To get the training teachers need, no other medium will deliver what AI can bring to the teaching profession. Teachers will learn skills to teach in a much more meaningful and fulfilling way each day, and manage their classrooms more effectively assisted by AI coaches. AI is best used and safest when it’s focused on the teacher. This advancement could lead to every child having a high-quality teacher, with a world of knowledge on teaching at their fingertips, no matter where they live.
