Use the following guiding questions to explore algorithmic bias and how it affects us and our world.
A search algorithm is the step-by-step procedure used to locate specific data among a collection of data.
Although they are often seen as neutral, they are very susceptible to reinforcing human biases.
Algorithmic Bias Introduction - Salem Press Encyclopedia of Science
Amazon’s hiring algorithm
Amazon`s algorithm used for hiring employees was found to be biased against women. Since most of the applicants were men, it was trained to favor men over women (Lauret, 2019).
Facebook ads
Facebook allowed its advertisers to intentionally target ads based on gender, race, and religion. For example, women were given priority in nursing or secretarial jobs in job advertisements, while job advertisements for gatekeepers and taxi drivers were mostly shown to men, especially men from ethnic minority backgrounds (Greig, 2021).
Healthcare risk algorithm
A health care risk prediction algorithm was designed to predict which patients may need additional medical care, but the incorrect results produced by the algorithm favored white patients over black patients (Simonite, 2019).
Zoom Video Communications:
Black and dark-skinned people have expressed frustration with Zoom's virtual background feature, which uses facial recognition technology to identify which parts of the screen should display the user and which should display the background image. Because the algorithms are supposedly unable to recognize faces of dark complexions properly, black and dark-skinned people appeared to disappear into their virtual backgrounds during Zoom calls (Costley, 2020).
Twitter algorithm- a racist
Users found that Twitter’s image cropping algorithm prefers to choose white people over black people. It was discovered that when images contained both former US President Barack Obama and Sen. Mitch McConnell, Twitter would consistently crop the shot just to show only McConnell (Collier, 2021).
Moreover, the algorithm was severely biased towards disabled people who used wheelchairs. Also, people who wore head coverings, which effectively cut out people who wore them for religious reasons, like a Muslim hijab (Collier, 2021).
Google search engine problem:
The google search engine show results with professional hairstyles at work vs unprofessional hairstyles at work where unprofessional hairstyles are all black women and professional hairstyles are white ladies. Thus, clearly representing racial discrimination. Similarly, from the Safiya Umoja Noble's Ted Talk, she complained that just by typing "why Black people are so...." into a search engine results in suggestions such as "so loud, so disrespectful, so ugly," and etc. Thus, explains that the google search engine algorithms have coded biases.
Societal Bias in Google
Google recently discovered that its advertising system enables advertisers to discriminate against nonbinary and transgender people. Those running advertisements on Google or Google-owned YouTube had the option of excluding those of "unknown gender," or those who hadn't identified themselves as male or female.
This essentially allows advertisers to discriminate against persons who identify as a gender other than male or female (whether intentionally or inadvertently), placing them in violation of federal anti-discrimination laws. Google has since changed its advertising settings. This is an example of algorithmic data bias being shaped by societal bias.
