Biased algorithms: the latest evolution of systemic racism

Category:  Opinions
Saturday, November 23rd, 2019 at 2:08 PM

Huge strides regarding racial equality have been made in the United States. Over the last few centuries, people of color have fought for freedom: the right to pursue life, liberty and happiness with the same vigor and support as their white counterparts.

 However, a country built on the backs of slaves is not easily changed. The United States was designed such that people of color were defaulted to second-class citizenship for hundreds of years, creating a society based inherently on inequity.

Some would argue that racial inequality is over and that evidence to the contrary furthers divisive “identity politics” and renders us all “snowflakes.” It is true that people of color are no longer enslaved and that legal protections now exist against hate speech and hate crimes. It is true that theoretically, a black person and a white person should have the same chance of success in America. Unfortunately, there are still inequalities to address in our society, and shying away from them only perpetuates those inequalities. 

People of color are disproportionately arrested and imprisoned compared to other demographics. As per the NAACP, African Americans are imprisoned at 5x the rate of white people. 

“If African Americans and Hispanics were incarcerated at the same rates as whites,” reads the NAACP website, “prison and jail populations would decline by almost 40%.” Furthermore, people of color receive harsher sentences for committing the same crimes as white people. “African Americans and whites use drugs at similar rates, but the imprisonment rate of African Americans for drug charges is almost 6 times that of whites.”

 This disparity is among the most obvious signs of continued structural racism in the United States.

According to the Glossary for Understanding the Dismantling Structural Racism/Promoting Racial Equity Analysis, published by The Aspen Institute, structural racism is defined as, “A system in which public policies, institutional practices, cultural representations, and other norms work in various, often reinforcing ways to perpetuate racial group inequity. It identifies dimensions of our history and culture that have allowed privileges associated with “whiteness” and disadvantages associated with “color” to endure and adapt over time.” 

The term structural racism is arguably synonymous with the term systemic racism, according to the Aspen Institute. The site reads: “If there is a difference between the terms, it can be said to exist in the fact that a structural racism analysis pays more attention to the historical, cultural and social-psychological aspects of our currently radicalized society.” These terms can be used interchangeably to relate the implications of a less obvious form of systemic racism: the biased algorithms now impacting communities of color. 

Studies show the negative impact of racial bias on the effectiveness of facial recognition technology used by law enforcement and self-driving cars, as well as an algorithm used to determine the need for health care follow-ups. 

It seems that as technological advancements further our society, so too do they promote the archaic notions of race embedded in our cultural subconscious. According to Clare Garvie and Jonathan Frankle of The Atlantic, a 2012 study conducted on mug-shots in Pinellas County, Florida to test the algorithms used by police in California, Maryland, Pennsylvania and others “found that all three algorithms consistently performed 5-to-10 percent worse on African Americans than on Caucasians. One algorithm, which failed to identify the right person in 1 out of 10 encounters with Caucasian subjects, failed nearly twice as often when the photo was of an African American.” 

As reported by Garvie and Frankle, African Americas are up to 2.5x more likely to be targeted by police surveillance than members of other races, according to some estimates. “This over-representation in both mug shot databases and surveillance photos will compound the impact of that 5-to-10 percent difference in accuracy rates. In other words, not only are African Americans more likely to be misidentified by a facial recognition system, they’re also more likely to be enrolled in those systems and be subject to their processing,” Garvie and Frankle wrote.

Facial recognition technology is marketed as highly effective, but according to Garvey and Frankle, claims of effectiveness are nearly impossible to verify. The unfortunate and terrifying truth they outline is that “the facial-recognition algorithms used by police are not required to undergo public or independent testing to determine accuracy or check for bias before being deployed on everyday citizens.” They continued, “More worrying still, the limited testing that has been done on these systems has uncovered a pattern of racial bias.”

Incarceration epidemic aside, these flaws in the algorithms that comprise facial recognition software can have dire physical consequences for people of color. In her article, “Self-driving cars may be more likely to hit you if you have dark skin,” Karen Hao of Technology Review summarized a study done by the Georgia Institute of Technology on the racial biases found in facial recognition software controlling self-driving cars. 

“The researchers tested eight image-recognition systems (each trained on a standard data set) against a large pool of pedestrian images. They divided the pedestrians into two groups for lighter and darker skin tones according to the Fitzpatrick skin type scale, a way of classifying human skin color.” She said.

 The results were unsavory. “The detection accuracy of the systems was found to be lower by an average of five percentage points for the group with darker skin. This held true even when controlling for time of day and obstructed view,” said Hao. 

After getting hit by a racially biased self-driving car, a person of color may still find herself at the mercy of algorithms when receiving medical care. Research on a widely used algorithm for follow-up care in hospitals revealed, “Overall, only 18% of the patients identified by the algorithm as needing more care were black, compared to about 82% of white patients.”

 If the algorithm were to reflect the true proportion of the sickest black and white patients, those figures should have been about 46% and 53%, respectively. According to Shraddha Chakradhar of Stat, who summarized the study, “Dissecting racial bias in an algorithm used to manage the health of populations,” published in Science in October 2019. The algorithm in question is used by health systems for over 100 million people across the United States. 

The authors of the study attempted to mitigate the bias of the algorithm by retraining it with the biological data of patients instead of the insurance claims data utilized by the original program. This revamp of the algorithm “found an 84% reduction in bias,” according to Chakradhar. She continued, “Previously, the algorithm was failing to account for a collective nearly 50,000 chronic conditions experienced by black patients. After rejiggering the algorithm, that number dropped to fewer than 8,000. The reduction in bias emphasized what many in the health technology field believe: Algorithms may only be as good as the data behind them.”

Unfortunately, the data behind them tends to be skewed. In his interview with Chakradhar, Sendhil Mullainathan, Professor of Computation and Behavioral Science at The University of Chicago’s Booth School of Business and co-author of the study on the health follow-up algorithm, said, “In general, these algorithms are built on data and those reflect systemic biases, and so won’t the algorithm also reflect the biases?” 

Indeed, the systemic racism inherent to American society perpetuates itself. 

The first step? Awareness. According to Garvey and Frankle, “the conditions in which an algorithm is created — particularly the racial makeup of its development team and test photo databases — can influence the accuracy of its results.” To change the initial conditions that allow bias to infect the algorithms, we must first recognize that a change is needed and act accordingly. 

Garvey and Frankle wrote: “Facial-recognition systems are more likely to either misidentify or fail to identify African Americans than other races, errors that could result in innocent citizens being marked as suspects in crimes. And though this technology is being rolled out by law enforcement across the country, little is being done to explore — or correct — for the bias.” This carelessness is exactly what allows structural racism to continue poisoning the free world.

“Structural racism is not something that a few people or institutions choose to practice,” according to the Aspen Institute. Rather, it is a racially motivated undertone that blankets the entirety of a culture. It is a safe bet to assume the designers of the biased algorithms mentioned herein are not racists; that does not negate the fact that their products further push racial biases and systemic racism. 

From whips and nooses, to police dogs and fire hoses, to handcuffs and health care, the means for the oppression of people of color are ever-changing. The only constant is the inability of the majority of Americans to recognize systemic racism for what it is.

Martin Luther King Jr. said: “He who passively accepts evil is as much involved in it as he who helps to perpetrate it. He who accepts evil without protesting against it is really cooperating with it.” I applaud the researchers who noticed, studied, and attempted to correct the racial biases found in algorithms used for facial recognition and health care follow ups. 

As a nation, we must constantly look at our habits, structures and processes with a critical eye to avoid complacency. 

Tags: opinions

View Our YouTube Channel
Edinboro TV
 
Find Us on Instagram