Public institutions now run on algorithms.
I am a computer scientist and public-interest technologist. I work inside government agencies to measure how algorithmic systems allocate scarce public resources, and I build the methods and tools that help institutions govern those decisions.
I wrote the foundational book in human-centered data science (MIT Press, 2022) to help the people who build data systems recognize the human choices behind the technical work.
I also write The Public Interest Technologist newsletter for the practitioners and regulators now deciding how AI enters public life.
I welcome inquiries about talks, advisory and consulting work, and media. I reply quickly to anything time-sensitive.
Release
There is no such thing as raw data. Every dataset is a series of human choices dressed up as objectivity.
Human-Centered Data Science
MIT Press · 2022
The first textbook in human-centered data science, now taught in graduate programs across North America, Europe, and Asia.
“We cannot engage in data science that doesn’t account for power. Histories and systems of race and gender must be taught to data scientists, because we know terrible wrongs can occur in the making and use of data. This book is a must-read to expose the next generation of data scientists to the consequences of their work.”
My lab works in four directions.
All of it inside institutions that are already running these systems.
My students and I do this work inside child welfare, housing, policing, health, and higher education.
Visit the lab ↗Critique
Auditing and theorizing the algorithmic systems already running inside public institutions.
Construction
Building alternatives, tools, and datasets with the communities and workers who use them.
Measurement
Testing whether those systems work, against criteria the affected communities set.
Governance
Shaping the rules that decide which systems exist, and under what oversight.
Work that governments and civil society draw on to govern AI.
Drawn on by the 2023 RMF Playbook, the United States' core federal resource for governing socio-technical AI risk.
Cited in the 2025 report's analysis of human oversight and accountability in public-sector AI.
Research used in the 2021 and 2024 reviews of Canada's federal Algorithmic Impact Assessment framework.
Advised the ACLU's audit of the Allegheny Family Screening Tool as its external technical expert.
Referenced in a multilateral analysis of AI's effects on the working lives of women.
Informs the Ministry of Higher Education's guide to responsible AI in colleges and universities.
Governments and regulators bring me in to help them govern AI.
Selected expert consultations and advisory roles across public agencies, regulators, and international bodies.
A regular voice on AI, data centres, and public accountability in Canadian media.
Coverage of my lab's study of Canada's federal AI register, which found some departments heavily dependent on American companies for their AI tools.
On the two data-centre motions before Toronto City Council, where land-use tools do not reach electricity demand.
Invited guest on the sovereignty and community stakes of Canada's AI data-centre buildout.
On Canada's national AI strategy and the public-trust gap it has to close.
On Hamilton's vote to pause new AI data centres. A pause is a reasonable answer when governance has not caught up.
Expert source on AI-altered listing photos, in a market where Canada requires no disclosure of altered images.
Open to talks, advisory and consulting, and press.
Here are the kinds of inquiry I welcome. A short, specific email is the best way in.
For academic, public-sector, and industry audiences, on AI governance, human-centered data science, and technology inside public institutions.
With public agencies, non-profits, and policy teams deploying, evaluating, or auditing AI and data systems.
Interviews, panels, and expert background for journalists covering AI, data centres, and accountability in public life.
A ready-to-use bio.
For journalists, event organizers, and anyone introducing me: choose a length and copy it. All three are current.
Shion Guha is an Assistant Professor of Information and Computer Science at the University of Toronto, where he directs the Human-Centered Data Science Lab and is Faculty Advisor on AI and Cities at the School of Cities. He studies how governments use AI to allocate public resources. He is the author of Human-Centered Data Science (MIT Press, 2022).
The second book is being written from the field.
For Public Interest Technology, I'm collecting anonymous field notes on the gap between tech policy and how it gets implemented inside public agencies. I read every submission.
Submissions are anonymous. Contributors can opt in to receive synthesized findings through The Public Interest Technologist.
Share a case study ↗I am recruiting Postdocs and PhD students.
If you care about public-interest technology, human-centered data science, or AI governance, and you want to do fieldwork inside public institutions, I would like to hear from you. I advise in the Faculty of Information and the Department of Computer Science.