All the world is talking about ChatGPT, but I am more interested in ChatAlgebra. I am interested in lowering the barriers to participation in computer science (CS) education, and increasing the diversity of the students who study CS and related disciplines. And I think ChatAlgebra may be a key that can unlock the gate.
The curricular structure of K-12 education devotes enormous emphasis (and time) to the goal of Algebra for All. Our educational institutions are not very successful at this goal; most students struggle in Algebra. Further, success in Algebra is inequitably distributed. The result of treating Algebra as a gateway course is to limit diversity of participation in computer science and STEM education. The ongoing pressure to improve Algebra scores is also a problem: the more schools double-down on Algebra, the less time there is to add CS and data science coursework to the curriculum. Thus, for a majority of students, the traditional Algebra sequence closes the gate to their opportunities to learn CS.
Revolutions can occur when an old structure no longer makes sense. In then wake of the interest in ChatGPT, we could recognize the availability of powerful systems that automatically process math symbols, such as Computer Algebra Systems (CAS). These have existed for decades — indeed, today a student can purchase a CAS calculator for less than $200. But use of CAS is not encouraged in K-12 education; generally, K-12 students are not taught to how computers could help them to do algebraic symbol processing, and instead studets are left to struggle to move around symbols in ways that computers could do for them. Does spending so much energy on Algebra for All still make sense, given that computers can do so much of this symbolic manipulation work for people?
Let’s define ChatAlgebra by analogy. One feature of the instantly-famous ChatGPT is that it can write the typical essays at the quality required of high school students. ChatAlgebra would be an AI-enabled interactive tool that can do all the calculations and symbolic transformations that appear on high school assessments, including the exams that serve as gatekeepers to CS courses and higher education. Although ChatAlgebra does not literally exist today, it could be an extension of CAS. What would happen if we changed the rules of math education — and especially math exams — so that students could use a powerful AI-enabled ChatAlgebra during their math learning and assessment experiences?
This really isn’t a radical idea. Starting in the 1990s, students have been able to use four-function calculators in classrooms and on exams. Then, when graphing calculators became common, more advanced calculators were allowed in classrooms and exams. There was hand-wringing at the time about what providing calculators would do to students’ math abilities, but little of that bad stuff occurred. Teachers know when to say "Students, calculators down now" -- so student still learn to do calculations and mental arithmetic. And when we are testing students, calculator use is limited on portions of the test, so those portions of the test remain valid. Historically, there have been few negative impacts of introducing calculators and graphing calculators to schools and tests. And there has been a well-documented positive impact: students make more progress in learning the conceptual aspects of mathematics and graphing when cognitive load related to doing calculations and plotting points is reduced.
By analogy, I believe that we could teach Algebra in less time and with greater success if we reduced the cognitive load now associated with doing symbolic manipulations by hand. I am NOT saying we should stop teaching Algebra; I am also not arguing that learning Algebra is unnecessary to learning CS. Even with a ChatAlgebra tool, a student would still have much to learn about how to use those symbolic manipulations in appropriate ways to solve meaningful problems; they would still need conceptual understanding of how and why Algebra works. Here's what I am saying: we are spending a lot of valuable curricular time to hone students' ability to accurately and quickly manipulate symbols, and we aren't enabling most students to succeed. Meanwhile, there’s an app for that. So let's rethink the traditional math course sequence that occupies so much time in school. Let's extend how we've already re-thought math in light of calculators, graphing calculators, spreadsheets, and more to also allow students to use ChatAlgebra as an aid to doing symbolic mathematical work.
Today, we spend about two years of middle school and two more years in high school trying (and often failing) to get every student to pass Algebra. If every student was enabled to use ChatAlgebra to power through error-prone symbolic manipulations, perhaps we could cut the curricular time for teaching Algebra.
By giving students access to symbolic manipulation tools, perhaps we could only teach Algebra for half as much time. Then we might be able to give every student two full years of well-planned curricular opportunities in computer science and data science. Students who hate Algebra may find CS and data science to be more motivating and approachable topics. And guess what? While they were studying CS and data science, they would also be learning more about Algebra, because the topics do go hand-in-hand.
We need a ChatAlgebra revolution to reduce the gatekeeping role of the traditional Algebra sequence, to ameliorate the student suffering associated with endlessly pushing symbols around, to overcome inequitable results of gatekeeping Algebra exams, and to free up some of the curricular space devoted to Algebra. Using the newly free time, we could enable a more diverse body of students to become engaged and proficient in computer science and data science skills.
Jeremy Roschelle is Executive Director of Learning Sciences Research at Digital Promise and a Fellow of the International Society of the Learning Sciences.