From the course: Introduction to Auditing AI Systems

Unlock the full course today

Join today to access over 24,600 courses taught by industry experts.

Ethics of scoring and classifying humans

Ethics of scoring and classifying humans

From the course: Introduction to Auditing AI Systems

Ethics of scoring and classifying humans

- [Instructor] The ethics of scoring and classifying humans has been a contentious issue in machine learning for some time. Machine learning is often described as applied statistics, which unfortunately has roots in some biased practices. The fathers of statistics, Francis Galton, Carl Pearson, and Ronald Fisher, were proud eugenicists who developed popular statistic methods to discriminate and improve the human gene pool through selective breeding and sterilization. As a product of statistics, AI often perpetuates unwanted patterns of data and raises concerns about the unfair treatment of marginalized groups. While modern approaches to scoring and classifying humans don't explicitly advocate for eugenic practices, they can perpetuate similar forms of bias, including digital phenology. An algorithm trained on data that reflects existing social inequalities will likely associate demographic characteristics with negative…

Contents