This extended show discusses the potential of bias in datasets, as well as the dangers of facial recognition.

The article describes Joy Buolamwini's research into how facial recognition software works much better for white men than women and people of color, why this happens, and the implications of this bias.

This article includes insight from three women in AI into the root causes of bias in AI and advice to reduce it.

In this article, Heidi Ledford explains how algorithms that allocate healthcare to patients have a racial bias. The article explains that because of systemic racism in healthcare, black patients were typically assigned lower risk scores -- making them less likely to receive personalized care.

This article describes how ImageNet Roulette, an app that labels faces based on Artificial Intelligence, shows biases in facial recognition software and how they can be based in the biases of the engineers themselves.

Canales reports on an internal change in the language used by Twitter in code and related topics. The article includes a list of the words and phrases that were changed, and their new alternatives.

This article, from the ACLU, provides insight into the impact of AI on racism in the criminal justice system, from tools used in law enforcement to those used to suggest bail amounts or whether someone should get parole, ending with a description of software to increase accountability..

Using the COVID-19 race tracker and introductory CS concepts, this project allows students to learn about the disproportionate effect of COVID-19 on Black Americans while practicing computer science concepts. The above link will take you to the main page for this project, which includes guiding questions, the original data, and an example solution. Teachers may give students some of the example solution to tailor the project, depending on the skills they wish to highlight. The assignment can be either take home or in class, but discussion is encouraged either during the project or after finishing it.

Apple uses machine learning for their autocorrect/suggestions, and their machine learning decided that a white male was the definition of a ceo, cfo, doctor, and other top positions. Their second option was a white female.

Wilson reports on a change that Nextdoor implemented in their user interface, to reduce racial profiling. Wilson further explains where the rationale for this decision came from, and the impact that it has had.

Learn about the history of computer science and how the field has become so unwelcoming to women.

This article, from Reuters, gives the example of bias in AI that Amazon uses for recruiting and how it specifically affects female candidates. The article explains multiple ways that this software has been shown to be biased and enumerates other companies that seem to be using similar algorithms.

In this piece, Querini decribes "Inclusive Design," or how to design with different people in mind. The author further describes how one can design for gender and what considerations to take into account when designing a system or user interface.

This article reports on an incident where a husband and wife were provided with vastly different credit lines from the Apple Card algorithm, despite sharing many of the same assets and filing taxes jointly. The article also explains the response from Apple, Goldman Sachs, and other involved actors.

This article describes criticisms of a recent Stanford study that attempted to guess whether individuals were gay or straight based on their faces, using AI. The article further explains the response of a co-author of the study and his motivations for publishing the study.

This report describes the importance of having queer researchers in the HCI (or Human-Computer Interaction) field. 

Cirillo et al. explain how current biomedical technologies that use AI models to tailor treatments can be biased based on gender. The paper outlines how current algorithms do not consider how gender plays into death and disease and in doing so, produces sub-optimal results. The paper outlines in depth current technologies and their effect on women, both positive and negative.

Listen to understand how systems can be changed for sustainable equity, as well as to learn about the organization CSforAll.

Krzyminski delves into the question of how to construct gender selection forms so that they are both inclusive and useful for the people collecting the data. The author provides four examples of gender selection forms and how they perform within each of these criteria.

This article explains how Artificial Intelligence algorithms can be biased against the LGBTQ+ community. The article touches on censorship of language used by this community, how AI can learn offensive languages and biases that it can then replicate, and many other issues.

This article describes how Nextdoor reduces racial profiling through a change in their UX and how the same principle can be applied throughout UX design. The article further explains similar UX decisions, spearheaded by social psychologist Jennifer Eberhardt.

Jean Ryoo, a researcher at the UCLA Computer Science Equity Project, speaks about equitable access to Computer Science Education.

Get inspired to create and have an impact on the world during our digital revolution.