Intersectional Accuracy Differences in Gender Classification
Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification
Recent studies demonstrate that machine learning algorithms can discriminate based on classes like race and gender. In this work, we present an approach to evaluate bias present in automated facial analysis algorithms and datasets with respect to phenotypic subgroups. Using the dermatologist approved Fitzpatrick Skin Type classification system, we characterize the gender and skin type distribution of two facial analysis benchmarks, IJB-A and Adience. We find that these datasets are overwhelmingly composed of lighter-skinned subjects (79.6% for IJB-A and 86.2% for Adience) and introduce a new facial analysis dataset which is balanced by gender and skin type. We evaluate 3 commercial gender classification systems using our dataset and show that darker-skinned females are the most misclassified group (with error rates of up to 34.7%). The maximum error rate for lighter-skinned males is 0.8%. The substantial disparities in the accuracy of classifying darker females, lighter females, darker males, and lighter males in gender classification systems require urgent attention if commercial companies are to build genuinely fair, transparent and accountable facial analysis algorithms.
Data preprocessing techniques for classification without discrimination
Recently, the following Discrimination-Aware Classification Problem was introduced: Suppose we are given training data that exhibit unlawful discrimination; e.g., toward sensitive attributes such as gender or ethnicity. The task is to learn a classifier that optimizes accuracy, but does not have this discrimination in its predictions on test data. This problem is relevant in many settings, such as when the data are generated by a biased decision process or when the sensitive attribute serves as a proxy for unobserved features. In this paper, we concentrate on the case with only one binary sensitive attribute and a two-class classification problem. We first study the theoretically optimal trade-off between accuracy and non-discrimination for pure classifiers. Then, we look at algorithmic solutions that preprocess the data to remove discrimination before a classifier is learned. We survey and extend our existing data preprocessing techniques, being suppression of the sensitive attribute, massaging the dataset by changing class labels, and reweighing or resampling the data to remove discrimination without relabeling instances. These preprocessing techniques have been implemented in a modified version of Weka and we present the results of experiments on real-life data.
Gender Obliviousness in Autism:
Accuracy Concerns Are Not Identity Concerns
When Good Algorithms Go Sexist: Why and How to Advance AI Gender Equity (SSIR)
Seven actions social change leaders and machine learning developers can take to build gender-smart artificial intelligence for a more just world. <meta property=
Sheryl Sandberg: The AI gender gap is about recognition
New Lean In survey says bosses aren't giving women credit for using AI.
Artificial Intelligence and gender equality | UN Women – Headquarters
The world has a gender equality problem, and Artificial Intelligence (AI) mirrors the gender bias in our society. Although globally more women are accessing the internet every year, in low-income countries, only 20 per cent are connected. The gender digital divide creates a data gap that is reflected in the gender bias in AI.
The Large Gender Gap in Who Uses AI
A recent study finds that AI usage tilts heavily toward men. Part of the reason: Women worried they might be penalized for using AI.
If AI is the future, gender equity is essential | NetHope
As the world faces many serious challenges — from climate change and poverty to displacement and social injustice – women and girls are key to tackling those issues and shaping future prosperity for all. Artificial Intelligence (AI), along with other digital tools, has the potential to help us address those challenges and open new opportunities. Seizing that potential requires us to take meaningful action to mitigate the risks AI poses to women and girls, and to use AI to advance gender equity.
The Interracial Cuck Porn Theory of Everything
Everything is gender, but to understand gender you must understand sex.
Not My Type | Stanford University Press
In the world of online dating, race-based discrimination is not only tolerated, but encouraged as part of a pervasive belief that it is simply a neutral, personal choice about one's romantic partner. Indeed, it is so much a part of our inherited wisdom about dating and romance that it actually directs the algorithmic infrastructures of most major online dating platforms, such that they openly reproduce racist and sexist hierarchies.
AI Is the Future—But Where Are the Women?
Just 12 percent of machine learning researchers are women—a worrying statistic for a field supposedly reshaping society.
AI’s Gender Gap - Haas News | UC Berkeley Haas
Generative AI could boost productivity and reduce inequality, but women across the world are at risk of missing out. A sweeping analysis by Nick Otis, PhD 25, Assistant Professor Solène Delecourt, and colleagues from Stanford and Harvard found that women were about 20% less likely than men to use tools such as ChatGPT, Claude, and […]
Facts and Figures 2023 - The gender digital divide
The percentage of men and women using the Internet.
The Gender Gap In AI Use—And What’s Driving It, According To Lean In Survey
The new data from Lean In reveal a gender gap in AI use, as well as in support and recognition for those who do use it.
Women portrayed as younger than men online, and AI amplifies the bias - Haas News | UC Berkeley Haas
In a sweeping study published today in Nature, researchers at UC Berkeley Haas, Stanford, and Oxford/Autonomy University documented extensive age and gender distortion across online media—and found that common algorithms are amplifying the bias.
