Who is Kate Crawford?
Kate Crawford is an Australian researcher and artist studying AI’s relationships with power, labour, history and the environment. She is a research professor of communication at USC Annenberg.[1]
Education, teaching and digital culture
Her academic career developed initially in Australia. The University of Sydney’s institutional history dates her appointment to 2003, while she was completing her doctorate. She created an online media production course combining critical analysis of the internet with writing, design and website development. [7]
In 2006 she presented ABC Television’s music series SET. She also worked as a journalist in Australia and the United States. [8]
Electronic music and artistic practice
She formed B(if)tek with Nicole Skeltys. The duo participated in Australia’s electronic music scene, performed at festivals and venues including the Sydney Opera House, and recorded with Julee Cruise. This activity forms an artistic career in its own right, alongside her later visual research. [9]
She co-founded the independent label Deluxe Mood Recordings. In 2008, while an associate professor at UNSW’s Journalism and Media Research Centre, she received the Australian Academy of the Humanities’ Max Crawford Medal. Adult Themes had received the 2006 Manning Clark House Cultural Award. [10]
Adult Themes and social transformation
Adult Themes: Rewriting the Rules of Adulthood was originally published in 2006. It was one of her early extended works of cultural analysis. [30]
The book questions whether marriage, home ownership and a stable career universally define maturity. It connects generational stereotypes with economic changes and different ways of living, examining adult life and its public representation. [8]
Big data, interpretation and rights
In Critical Questions for Big Data, published with danah boyd in 2012, she argued that having more data does not remove problems of sampling, context and interpretation. The authors distinguished information volume from representativeness: an enormous collection of digital activity can still exclude populations or distort their practices. [11]
With Jason Schultz she published Big Data and Due Process in 2014. They examined how predictive profiles can affect individuals even when based on mistaken inferences. They proposed procedural safeguards for understanding and challenging data use in private decisions, offering an academic response to shortcomings in existing protections. [12]
AI Now, ENS and USC
Crawford and Meredith Whittaker co-chaired the 2016 AI Now symposium organized by NYU’s Information Law Institute and the White House. Researchers, public officials and practitioners met to examine near-term AI effects on inequality, health, labor and ethics. [13]
They founded the AI Now Institute, launched at NYU in November 2017. It brought together interdisciplinary research on civil rights, biased data, infrastructure and automation. Its establishment turned a series of discussions and meetings into an institutional program studying technology’s social effects. [14]
In 2019 she became the inaugural holder of the Abeona-ENS visiting chair in Artificial Intelligence and Justice in Paris. The initiative connected technical research with questions of justice, bias and fairness. This was a historical appointment within a chair that hosts visiting specialists. [15]
She joined USC Annenberg in March 2021 as a research professor of communication and science and technology studies. Her appointment strengthened research on sociotechnical systems: technologies examined alongside the social practices, rules and power relationships within which they operate. [16]
Atlas of AI: a material interpretation of technology
Yale University Press published Atlas of AI in 2021. It examines natural resources, infrastructure, data and human labor as components of artificial intelligence. Crawford offers a critical interpretation of the industry as an extractive activity, connecting computer systems to the material costs and decisions about power that sustain their operation. [3]
Anatomy of an AI System
With Vladan Joler she created Anatomy of an AI System, a 2018 work acquired by New York’s Museum of Modern Art. Combining research with graphic representation, it exemplifies the museum’s use as a place to examine technological infrastructures that usually remain out of sight. [20]
Excavating AI, Training Humans and the debate over images
In their 2019 essay Excavating AI, Crawford and Trevor Paglen examined categories and labels in image datasets used for machine learning. They asked which concepts become classes, how images are assigned to them and what assumptions about people enter that process. [17]
Training Humans, conceived by both, ran at Osservatorio Fondazione Prada from September 2019 to February 2020. The exhibition traced training-image collections from the 1960s onwards. Displaying them as historical objects brought visual classification and the decisions behind it before the public. [18]
The project also attracted criticism. In 2021 Michael J. Lyons, a co-creator of JAFFE, challenged Crawford and Paglen’s description of its images and stated that they had exhibited them without seeking permission. His paper presents one dataset creator’s position; it documents a professional dispute, rather than a court ruling. [19]
Calculating Empires and the history of technology
During her Cybernetic Imagination residency at the Australian National University in 2022–2023, she worked with Joler on Calculating Empires. It premiered at Fondazione Prada in November 2023. The project extended visual research from a single device to a history of relationships between technology and power. [21]
The installation traces five centuries from 1500 in a twenty-four-meter visualization. It connects changes in communication and calculation with colonialism, militarization and automation. Crawford and Joler are credited alongside a research, design and production team whose collaboration makes this documentary scale possible. [22]
Calculating Empires received the S+T+ARTS Grand Prize for Artistic Exploration in 2024, within the European Commission’s awards organized by Ars Electronica. [23]
In 2025 it received the Silver Lion at the Venice Architecture Biennale, where it was presented in partnership with Powerhouse Museum. The recognition was awarded to Crawford and Joler’s joint work. [24]
Knowing Machines and dataset research
Crawford is lead principal investigator of Knowing Machines, an interdisciplinary project examining how data used to teach models to interpret the world are constructed and employed. Its team brings together technology research, history, art and law to study the decisions that turn a collection into a training reference. [25]
In 2023 she co-edited with Mike Ananny A Critical Field Guide for Working with Machine Learning Datasets, written by Sarah Ciston. It examines dataset selection, documentation, use and lifecycle, asking practical questions about provenance, purpose and responsibility instead of treating data as interchangeable material. [26]
In Building Better Datasets, published with Will Orr in 2024, she interviewed eighteen dataset creators. The study offered seven recommendations concerning quality, documentation, privacy, consent and unintended uses. Its basis is qualitative: it draws on professional experience to improve practices, without presenting a universal measurement of all projects. [27]
Recent activity
Her 2026 professional biography identifies USC and Knowing Machines as central to her work and records her appointment as an Amherst College Presidential Scholar. It also notes her role in co-founding FATE, Microsoft’s research group on fairness, accountability, transparency and ethics in AI. [28]
Between August and September 2026 she held a Distinguished Fellowship at Berlin’s Weizenbaum Institute. The program included work on a new book and exchanges with specialists and the public about AI’s social and environmental consequences. It was a temporary research stay, rather than a permanent university move. [29]
Workplace surveillance and health-data privacy
In Limitless Worker Surveillance (2017), Crawford, Ifeoma Ajunwa and Jason Schultz examined productivity applications and employer wellness programs collecting health information. They analyzed shortcomings in US legal protection and proposed safeguards against surveillance extending beyond work. [6]
For OSH, this raises a specific decision when introducing health sensors or applications: separating preventive purposes from performance monitoring. An alert needed to correct an exposure does not justify turning all of a worker’s personal information into a productivity indicator.
The labor chain behind AI
In Anatomy of an AI System, Crawford and Vladan Joler analyze a connected device through its materials, data and human labor. Their map includes mineral extraction, manufacturing, logistics, data labeling and electronic waste, showing that AI depends on activities with very different working conditions. [4]
Prevention should examine each activity and group in that chain, alongside the end user’s privacy. Both works provide criteria for investigating conditions and responsibilities; their analyses do not constitute a current exposure assessment of every supplier.
