Bias in algorithms: A Mirror to Our Own Prejudices

Bias in Algorithms: Shedding Light on Our Own Biases

Understanding Algorithmic Bias

Algorithms were created with the intention of streamlining processes and making decisions more impartial. They were meant to assist in various aspects of our lives, from hiring processes to judicial decisions and healthcare allocation. However, as time has passed, it has become evident that algorithms can harbor biases just like the human decision-makers they are designed to support. In a study conducted by Carey Morewedge, a professor at Boston University, it was discovered that individuals are more adept at recognizing biases in algorithmic decisions compared to their own, even when the decisions are identical. This research, published in the Proceedings of the National Academy of Sciences, proposes a novel perspective on how acknowledging biases in algorithms can potentially assist human decision-makers in identifying and rectifying their own biases.

Unveiling Structural Biases

The study highlighted how algorithms can inadvertently perpetuate biases present in the human decisions used to train them. For example, Amazon’s hiring algorithm in 2015 demonstrated a gender bias by favoring male applicants over female ones. However, this bias was merely a reflection of the existing bias within Amazon’s hiring practices. Morewedge emphasizes that while algorithms have the capacity to magnify human biases, they also serve as a mirror reflecting the structural biases prevalent in our society. By analyzing a multitude of decisions across various individuals, algorithms can unveil underlying biases that may not be apparent at an individual level, shedding light on systemic issues within organizations and systems.

Related Video

Recognizing Bias in Others and Ourselves

The study conducted by Morewedge and his team involved a series of experiments aimed at assessing social biases such as racism, sexism, and ageism. Participants were tasked with evaluating algorithmic decisions and their own decisions, with some decisions falsely attributed to algorithms. The results demonstrated that individuals were more inclined to identify bias in decisions they believed originated from algorithms rather than in their own decisions. This phenomenon, known as the bias blind spot, indicates that people are better at recognizing bias in others than in themselves. The research further revealed that individuals were more likely to correct for bias in decisions attributed to algorithms post hoc, underscoring the importance of acknowledging bias to mitigate its impact in the future.

Algorithms as Tools for Self-Reflection

Morewedge’s research showcases how algorithms can serve as tools for introspection, allowing individuals to confront their biases and take steps towards rectifying them. By presenting algorithms as mirrors that reflect our biases back to us, the study suggests that algorithms can play a pivotal role in helping individuals become more aware of their own biases. Rather than solely focusing on developing methods to reduce bias in algorithms, Morewedge advocates for a dual approach that involves enhancing algorithms while simultaneously working towards reducing biases within ourselves. The study posits that algorithms, despite their potential to amplify biases, can also be instrumental in fostering self-awareness and facilitating personal growth by enabling individuals to recognize and rectify their biases.

The research by Morewedge offers a fresh perspective on the role of algorithms in highlighting biases, both in the systems they operate within and in the individuals they interact with. By leveraging algorithms as reflective tools, individuals can gain valuable insights into their biases and work towards creating a more equitable and unbiased society. The study underscores the dual nature of algorithms as both amplifiers of biases and catalysts for self-improvement, emphasizing the importance of utilizing technology to not only enhance efficiency but also promote introspection and personal growth in combating bias.

Links to additional Resources:

1. Technology Review: Algorithms Can Be Biased. Here’s How to Fix Them 2. Wired: Can the Bias in Algorithms Help Us See Our Own? 3. The New York Times: Algorithms Can Be Biased, and That Can Lead to Discrimination. 

Related Wikipedia Articles

Topics: Algorithmic bias, Bias blind spot, Carey Morewedge

Algorithmic bias
Algorithmic bias describes systematic and repeatable errors in a computer system that create "unfair" outcomes, such as "privileging" one category over another in ways different from the intended function of the algorithm. Bias can emerge from many factors, including but not limited to the design of the algorithm or the...
Read more: Algorithmic bias

Bias blind spot
The bias blind spot is the cognitive bias of recognizing the impact of biases on the judgment of others, while failing to see the impact of biases on one's own judgment. The term was created by Emily Pronin, a social psychologist from Princeton University's Department of Psychology, with colleagues Daniel...
Read more: Bias blind spot

Negativity bias
The negativity bias, also known as the negativity effect, is a cognitive bias that, even when positive or neutral things of equal intensity occur, things of a more negative nature (e.g. unpleasant thoughts, emotions, or social interactions; harmful/traumatic events) have a greater effect on one's psychological state and processes than...
Read more: Negativity bias