By Grace Johnson, Mar. 9, 2021
Using their extensive knowledge in philosophy, education and artificial intelligence, professors collaborated in early 2019 to begin their current project, Supplemental AI-Assisted Learning System, or SAILS, aspiring to eliminate socioeconomic and racial inequities in STEM majors.
The inspiration for the project emerged from Cal Poly Pomona philosophy professor Alex Madva and Cal State Fullerton computer engineering professor Yu Bai’s recognition of the structural challenges and barriers that impede STEM students of color, resulting in lower retention rates.
“Similar to STEM, the field of philosophy is a historically exclusionary field that still has too many white, able-bodied straight men, coming from relatively privileged backgrounds,” said Madva. “I’m interested in and committed to taking steps toward making my field more inclusive and I think that the solutions for this are going to have much in common with solutions for similar problems in STEM.”
The team’s goal is for the AI program to use student data to make recommendations to students and advisers about which classes to take and direct students to resources and potential support. Possible student data includes course grades, course names, classwork, discussion board posts and possibly more personal information such as ethnicity, race or economic standing, if the individual is willing. This programming would then be available for advisors, counselors, professors and students to use.
Before any AI development can begin, Bai and Madva, in partnership with a cohort from the University of Missouri-Kansas City, must come together to begin gathering data, which requires funding they have not yet obtained. The group applied for a $350,000 National Science Foundation grant last October and expects to receive results in April.
“If we want to make change in the world, we must have data, and that takes money,” explained Madva. “There is a larger body of research across the country that we are slowly building on which looks at the factors making certain fields of study more or less inclusive. Although we have some access to CSU’s information via dashboard, we want to be able to see the ethnic achievement gaps with different majors of schools across the nation.”
Extracting data, the team will be able to further conceptualize what kind of instruction students have received from advisors and compare it with participants’ cultural identities. According to the team’s theory, some students may have been steered by an advisor toward a class that others were not, based on socioeconomic status or race. Bai and Madva’s main concern with this reality is that some students are not being encouraged to take on STEM classes, while other students are.
The main goal of SAILS is to halt socioeconomic biases in education and to give students an equitable chance at success.
“Different groups in campus populations are facing challenges, but with technology and artificial intelligence development, we can help,” said Bai.
Currently, Bai is utilizing his classroom to pilot confined versions of the program. During the fall semester of 2020, Bai presented some of these ideas to his students at Fullerton and implemented an AI program to identify students’ areas of difficulty on midterms. With this research, Bai identified similarities among students’ struggles and mitigated the issues by reteaching and reextending the exam to the classes. This led to a 20% increase in the class’ success rate, in regards to grades.
According to Madva, programs such as this can help pinpoint barriers that students face and can be trained to detect ethnic and gender discrimination in the educational system to provide possible solutions for the future.
“I think there is potential for students to understand how to study, where to focus, how much time and when to do so,” said University of Missouri-Kansas City School of Education Associate Research Professor Alexis Petri. “There are so many myths and misconceptions about STEM degrees that prohibit many students from pursuing something that might be difficult but will lead to a career that is rewarding.”
The three campuses collaborating on this research host students who have been historically excluded from STEM majors.
At CPP, 49% of the student body is Latinx, 21% Asian, 3% African American and less than 1% other demographics. At CSUF, 44% of the student body is Latinx, 2% African American, 21% Asian and 7% other. At the University of Missouri-Kansas City, 12.4% of the student body is African American, 8.2% Latinx, and less than 1% for each of the other demographics.
This three-way partnership has helped reach a more diverse pool of data because each university differs demographically.
“At the University of Missouri-Kansas City, we serve a large percentage of first-generation college students,” explained Petri. “What excites me most about this collaboration is the idea that we can find ways to bridge emerging technologies, such as AI, with teaching and learning. After all, one of my greatest concerns is that students won’t understand the financial impact of dropping classes or leaving university in debt without a degree.”
With ongoing discovery of biases in the educational system, Madva stresses that no matter how much research and preparation it takes, the group will continue to root out the source of these challenges.
“The most promising thing about artificial intelligence is that it can absorb so much more data than humans, who can only focus on select variables at a time,” stated Madva. “If we are able to learn more about students, we are able to give personalized feedback to them which will hopefully lead to the creation of tools for both students and advisors to use. We hope that SAILS will bring awareness to the obstacles that different students are facing.”
Feature image courtesy of Markus Spiske.