Meet the Faculty: Emily Black
This entry is part of our Meet the Faculty blog series, which introduces and highlights faculty who have recently joined CDS.
Meet Emily Black, who is joining CDS this fall as Assistant Professor of Computer Science, Engineering, and Data Science. Black brings her expertise in responsible AI, algorithmic fairness, and technology policy to address critical challenges at the intersection of machine learning and societal impact. Her work combines advanced machine learning techniques with a nuanced understanding of legal and policy frameworks to detect and mitigate bias in AI systems used in high-stakes societal domains.
“Increasingly, rules governing huge decisions in people’s lives are not just made in government, but at tech companies and in academia — or anywhere where predictive algorithms are created and reasoned about,” Black explains. “But these algorithms do not always work in an equitable fashion, and can bring about real harm, from racially biased healthcare distribution systems under-allocating resources to Black patients to tax audit selection systems over-selecting low-income individuals.”
Black’s research focuses on detecting and mitigating bias in AI systems used in high-stakes societal domains, with an emphasis on creating approaches that are complementary to relevant laws and regulation. Her work takes a holistic, “pipeline-aware” approach by examining the entire process of developing and deploying AI systems to uncover sources of bias that may not be apparent from traditional fairness metrics alone.
One area of Black’s research explores the phenomenon of model multiplicity — where many equally accurate models exist for the same problem — and its implications for fairness and anti-discrimination law. Her recent paper, “Less Discriminatory Algorithms,” co-authored with legal and science and technology studies scholars, won the Privacy Papers for Policymakers Award from the Future of Privacy Forum. This work proposes that entities using algorithmic systems in domains like housing, employment, and credit should have a duty to search for and implement less discriminatory algorithms (LDAs).
“A problem I’ve begun to pay attention to is that many of the debiasing techniques suggested in the computer science literature do not take legal restrictions on the use of demographic data into account, and thus can be difficult to use in high-stakes domains such as credit, housing, and employment,” Black notes. Her current focus is on developing AI debiasing techniques that are compliant with US anti-discrimination and privacy laws, while also working on effective policies to ensure these techniques are implemented in practice.
Black’s commitment to real-world impact is evident in her collaborations with government agencies and NGOs. In a study with the Internal Revenue Service, she found that changing an audit selection model’s prediction target was effective at reducing burden on low-income taxpayers while creating an increase in revenue for the IRS.
Looking ahead, Black plans to continue developing legally compliant techniques for reducing algorithmic discrimination by operationalizing her pipeline-aware approach. She also aims to extend her work to large language models and other generative AI systems as they become integrated into regulated decision-making processes.
“I am committed to ensuring that effective AI harm reduction does not only happen within academia, but in real-world deployed systems,” Black says. By serving as a technical bridge between government, industry, and academia, she hopes to develop highly implementable regulatory and technical tools to combat algorithmic discrimination in areas protected by civil rights law.
Black’s interdisciplinary approach and technical expertise make her a valuable addition to the CDS faculty, promising to advance the field of responsible AI and shape the future of technology policy.
By Stephen Thomas
