Would Teacherbots Eliminate Implicit Bias in Online Teaching?

By Dr. Kandis Boyd Wyatt, PMP, CLTD
Faculty Member, Transportation and Logistics, American Public University

Last month, I had the pleasure of moderating the third-ever APUS School of Business Online Debate entitled, “Will AI Replace Faculty in Online Schools?” With over 50 people in attendance, it was the largest debate to date and focused on infusing technology into the online classroom.

Start a degree program at American Public University.

The debate touched on many aspects of online learning, especially the topic of teacherbots, or automated assistants. Specifically, it addressed the following questions:

• Could a teacherbot do my job better than a human?
• Would students prefer to interact with a teacherbot instead of an instructor?
• What cost savings would teacherbots bring to the university?
• Could a teacherbot be more fair and unbiased than a human being?

Naturally as a full-time professor, this topic struck a chord with me on many levels and produced many questions, especially when it concerned implicit bias in online teaching.

What Is Implicit Bias?

The term “implicit bias” was coined 25 years ago by social psychologist Anthony Greenwald of the University of Washington. His research exposed an uncomfortable aspect of the human mind: “People have deep-seated biases of which they are completely unaware. And these hidden attitudes — known as implicit bias — influence the way we act toward each other, often with unintended discriminatory consequences.” As a result, implicit bias has infused itself into every aspect of life. Many people avoid this subject altogether because it is uncomfortable. Implicit bias involves self-reflection and in many cases analyzing how we contribute to this problem. Whether intentionally or not, we have all made judgments about others. However, the recent global outcry to address the social injustices of racism, sexism, homophobia, and classism has forced us to face the unpleasantness of implicit bias, pull back the curtain on it, and address the issue frankly. According to behavioral scientist Angela Duckworth, there are three underappreciated psychological facts that are often misunderstood about implicit bias:

1. Implicit bias is pervasive but not inevitable.
2. Implicit bias is subtle but not always unconscious.
3. Racial inequality goes far beyond implicit bias.

So while challenging to identify, implicit bias can be corrected and is the first of many layers of social inequities and injustices. From health to law to science, implicit bias affects us all in a multitude of ways. It creates a “good ol’ boy” system, where like-minded individuals are mentored, praised and rewarded. As a result, there is wage inequality, equitable education, criminal injustice, biased research, and redlining (the systematic denial of various services by government agencies as well as the private sector either directly or through the selective raising of prices). These problems further exacerbates the divide between the haves and have-nots.

We All Have Implicit Bias

Writing in Medical Economics, Dr. Rebecca Bernard highlights the fact “that every single one of us has a certain degree of implicit bias, or subconscious automatic thoughts about others that stem from our upbringing and social framing. Having these thoughts does not mean that we are inherently bad and evil.” But if we fail to identify areas of bias, blind spots in our knowledge and understanding of other groups of people, it can have detrimental effects. Implicit bias is one component in the broader system of historical, cultural and structural inequalities that perpetuates inequities in American society. One of the reasons why we fail to examine our biases is due to the emotions they produce. These emotions can be immensely powerful. And without help in processing them, we may develop maladaptive coping mechanisms, including denial or anger.

How to Address Implicit Bias

Implicit bias is often categorized as stereotyping. According to Science News, accurately measuring stereotyping is a challenge: “People are often reluctant to admit they have stereotypes, so asking directly about these beliefs is unlikely to provide an accurate measure of whether they endorse the [unpopular] ideas.” People are often unaware of their biases; feedback about implicit bias is often met with surprise and defensiveness. Understanding how biases evolve into systemic injustice is crucial in each and every facet of life. We must first agree that we all have implicit biases, and that does not make us a bad person. Next, start the conversation. By speaking with others we can acknowledge underlying issues and develop strategies to address them. Through allyship, which is defined as “an association with the members of a marginalized or mistreated group to which one does not belong,” each of us can acknowledge our desire to be a resource champion. Whether it is in the community or in the workplace, speaking, writing, and discussing these issues can educate us all.

Bias in the Classroom

I’d be remiss if I didn’t acknowledge how writing this article made me turn inward and re-examine my role as an online professor. So if you are an instructor in the online environment, these questions are a great way to start the conversation:

1. How does written communication influence your decisions? The vast majority of an instructor’s interactions with students are non-verbal. So on one hand this reduces biases related to diction and accent, but can increase biases when the written communication does not align with the instructor’s concept of “good grammar.”
2. Do instructors resonate with students from similar backgrounds? Students get an opportunity to introduce themselves and write a summary of their background during the first week. That can serve as a way to connect with fellow colleagues and the instructor. Could instructors have an implicit bias in favor of students who have similar backgrounds, higher military ranks, career achievements or common interests?
3. Do student photos affect student success? Students are allowed to post a picture of themselves that appears when they post forum discussions (it does not appear for assignments and one-on-one messages). About half of the students opt to use a picture, and the pictures vary. Some are professional pictures, some include a favorite pastime, and some include one’s family or pet. Does this provide a biased perception of the student? And for the students that do not provide a picture, are they more likely to be evaluated fairly?
4. Do past negative experiences with a student affect future success? If you are an online instructor in a specific program, chances are you are going to teach multiple courses and thereby have the opportunity to teach a student more than once.
5. Are student requests subject to implicit bias? Extensions are evaluated on a case-by- case basis at the discretion of the instructor. Do instructors use the same evaluation process for every student, or is implicit bias evident when some requests are approved and others are denied?
6. Do students develop implicit bias when interacting with instructors or fellow classmates based on their past experiences? Is it safe to say all students have implicit bias? How does this characteristic affect group work, class discussions and one-on-one interactions in the online environment?
7. Likewise, in what ways are graduate students evaluated differently from undergraduate students? Is there a higher expectation for graduate students because they have already obtained a bachelor’s degree?
8. And the most dreaded question of all, would a teacherbot remove these aforementioned biases and grade more fairly because of the absence of implicit bias? Could a teacherbot be trained to identify students exhibiting implicit bias and address it promptly? Or would the biases of the programmer be transferred to the AI technology?

The solution is to address implicit bias in the classroom head-on by acknowledging blind spots, conducting independent research and providing recommendations for change. By fostering change, it won’t matter if the lesson is taught by a human being or a teacherbot because the same expectations of implicit bias will be acknowledged and (hopefully) addressed fairly and uniformly. Let’s start the conversation TODAY.

About the Author

Dr. Kandis Y. Boyd Wyatt, PMP, is a professor at American Public University and has 20 years of experience managing projects that specialize in supply chain management. She holds a B.S. in meteorology and an M.S. in meteorology and water resources from Iowa State University, as well as a D.P.A. in public administration from Nova Southeastern University.