AI governance · HCI · Privacy-preserving systems
I research how AI systems can remain useful, accountable and human-centred when they move from models into real institutions.
My work spans AI assurance, privacy-preserving machine learning, human-computer interaction and the data governance of learning systems, across research, international policy and deployed software.
Incoming Postdoctoral Researcher, Umeå University, Department of Computing Science · from
Founder, AffectLog
Previously UNESCO, 2021-2025
Portrait to follow
Context
UNICEF Innocenti
Current
Expert advisory group · children's data governance
JA Institute
Current
Advisor · Research Advisory Council
Umeå University
from September 2026 · incoming
Postdoctoral research · computing science
AffectLog
2025-
Founder · AI assurance research and tooling
Digital Skills and Jobs Platform (European Commission)
AI in Health: how smart technology is breaking new frontiers in medical care (deep-dive)
Deep-dive article authored by Roy Saurabh and published on the European Commission's Digital Skills and Jobs Platform, which identifies him as founder of AffectLog, working on trustworthy AI governance and assurance for regulated sectors including healthcare.
ORCID
ORCID record 0000-0003-3439-7731
Canonical researcher identifier and the list of registered works. Employment and biography entries on ORCID are self-asserted.
UNESCO
2021-2025
2018-2025 · teacher development, digital learning and data governance
UNESCO
Digital Learning Week 2023, provisional programme
Lists 'Mr Roy Saurabh, Senior Project Officer, Section for Teacher Development, UNESCO' as a discussant on the Norms and Standards panel, 4-7 September 2023.
Broadband Commission for Sustainable Development
Working Group on Data for Learning
States that the UNESCO team coordinating the working group, led by Borhene Chakroun, acted 'with support from Juliette Norrmén-Smith, Associate Project Officer, and Saurabh Roy, Senior Project Officer'.
UNESCO-ICHEI
Delegation from UNESCO-ICHEI visits UNESCO headquarters for bilateral talks
Names him in Chinese as 联合国教科文组织AI研究高级项目官员Roy Saurabh先生, Senior Project Officer for AI research, UNESCO, in a bilateral meeting of 5 September 2024 with UNESCO's Division for Policies and Lifelong Learning Systems.
ORCID
ORCID record 0000-0003-3439-7731
Canonical researcher identifier and the list of registered works. Employment and biography entries on ORCID are self-asserted.
The through-line
Most of my work sits in the movement between these three, rather than inside any one of them. A finding that institutions cannot act on, and a requirement no system can implement, both fail in the same way.
What can we know?
What evidence and safeguards should institutions require?
How do those requirements become operational?
Current research
Claims about an AI system should be checkable by someone who was not involved in building it.
What evidence about an AI system is adequate for a particular decision, held by a particular role, in a particular deployment?
Institutions that cannot pool sensitive data can still learn together, if the governance is built alongside the method.
What has to hold, technically and institutionally, for federated analysis across sensitive datasets to be trustworthy rather than merely decentralised?
Systems that adapt to people need to be honest about how little they know about the person in front of them.
How can adaptive and affect-aware systems support learning and wellbeing without overreading sparse, intimate signals?
The governance of education and children's data decides what the technology is allowed to become.
How should public institutions govern data about people who cannot meaningfully consent, and what does that require of the systems built on it?
Selected work
Six items, each linked to the institution, registry or document that supports it.
Policy2025Author
Argues that governments can use sensitive education data without centralising it, and what governance that requires.
UNICEF Innocenti, Office of Strategy and Evidence
UNICEF Innocenti
Federated learning for children's data: encoding trust in education systems
Article authored by Roy Saurabh and published by UNICEF Innocenti on 19 May 2025, listed among the outputs of the good-governance-of-children's-data project.
UNICEF Innocenti
Good governance of children's data, advisory groups
Names Roy Saurabh among the current expert advisory group members shaping the second phase of UNICEF Innocenti's children's data governance project.
Publications2024Author
Doctoral research on what wearable sensor data can and cannot support when the question is clinical risk.
Université Paris Cité
theses.fr
Thesis record s310267
Doctorate awarded by Université Paris Cité, defended 17 December 2024, in Mathematics and Computer Science, supervised by François Taddei and Harri Ketamo.
ABES
Thesis DOI record
Registers the thesis against ORCID 0000-0003-3439-7731 with affiliation Université Paris Cité, confirming that the indexed author name 'Saurabh Roy' is the same researcher.
Policy2026Author
A deep-dive for the European Commission's skills platform on where clinical AI is genuinely useful, and what assurance regulated care requires of it.
Digital Skills and Jobs Platform, European Commission, DG CNECT
Research software2026Author
A reproducible public-records method for analysing AI risk as a function of deployment context rather than of the model alone.
Reports2023Writing team
Broadband Commission working-group report on the governance of education data, written by the UNESCO team supporting the group.
Broadband Commission for Sustainable Development, ITU/UNESCO
Broadband Commission for Sustainable Development
The Transformative Potential of Data for Learning
Acknowledgements name the UNESCO writing team as Juliette Norrmén-Smith and Saurabh Roy, with contributions from Carmela Salzano, Patrick John Devaney and Wayne Holmes.
Broadband Commission for Sustainable Development
Working Group on Data for Learning
States that the UNESCO team coordinating the working group, led by Borhene Chakroun, acted 'with support from Juliette Norrmén-Smith, Associate Project Officer, and Saurabh Roy, Senior Project Officer'.
Research software2026Author
Asks whether audit evidence is adequate for the specific decision, and the specific role, relying on it.
Policy & advisory
Advisory roles are advisory roles. None of these is employment at the institution named, and each links to the institution's own record of it.
Expert advisory groupCurrent
Office of Strategy and Evidence
Expert advisory group member, Good governance of children's data
One of around twenty members advising the second phase of UNICEF Innocenti's children's data governance project, which covers data governance in education technology, a fair data economy for children, and innovations in data governance.
AdvisorCurrent
Advisor, Research Advisory Council
Research Advisory Council member, advising on learning ecosystems and public-interest technology.
Publications
2026
SSRN
Preprint on how the choice of background distribution alters federated attribution results, and what that does to an audit.
PublicationsAuthor
2024
Université Paris Cité
Doctoral research on what wearable sensor data can and cannot support when the question is clinical risk.
PublicationsAuthor
2019
2019 IEEE 18th International Conference on Cognitive Informatics & Cognitive Computing (ICCI*CC)
Early work connecting cognitive science and socio-emotional learning to digital pedagogy.
PublicationsCo-author
Software
2026
A reproducible public-records method for analysing AI risk as a function of deployment context rather than of the model alone.
Research softwareAuthor
2026
A benchmark for reasoning over regulatory obligations as executable statements rather than prose.
Research softwareAuthor
2026
Asks whether audit evidence is adequate for the specific decision, and the specific role, relying on it.
Research softwareAuthor
Trajectory
Read as a career this looks like several. Read as a line of enquiry it is one, moving steadily closer to the institutions where these systems actually land.
2005-2014
Systems
Enterprise systems engineering and then consulting. The work was implementation, at a scale where a technical decision becomes an organisational fact.
2014-2018
Systems
Founding and technology leadership across skills and education platforms, including national skill-development programme technology. First sustained exposure to public-sector digital delivery.
2018-2020
Systems
Chief Technology Officer at UNESCO MGIEP, building learning technology with machine learning and socio-emotional learning at its centre. The question shifted from whether a system could be built to what it should be modelling.
2020-2021
Research
Lead data scientist at the Centre de Recherches Interdisciplinaires. Working directly on intimate data made the privacy constraints a research problem rather than a compliance step.
2021-2025
Governance
Senior Project Officer in UNESCO's Section for Teacher Development: teacher development, digital learning platforms and education data governance, including the UNESCO writing team for the Broadband Commission's data for learning reports.
2024
Research
PhD in mathematics and computer science at Université Paris Cité, on machine learning from wearable data for mental-health monitoring of healthcare professionals.
2025-
Governance
Founding AffectLog, and building reproducible tools for AI assurance. Advisory work with UNICEF Innocenti on children's data governance keeps the research attached to a population that bears the cost when governance fails.
2026-
Research
Postdoctoral research at the Department of Computing Science, Umeå University, from September 2026.
Talks
Invited lecture10 April 2025
CITIC, Universidade da Coruña · A Coruña, Spain
On handling sensitive data in health and education: where federated approaches help, and what governance has to be in place for them to mean anything.
Panel moderator17 October 2024
Ninth community workshop on Explainable AI, European Education and Culture Executive Agency (EACEA) · Brussels, Belgium
Moderated a panel on AI literacy and critical thinking, with Simona Petkova (DG EAC) and Romina Cachia (JRC, European Commission).
Writing
AI governance18 August 2026Digital Skills and Jobs Platform
Where clinical AI is genuinely useful, where it is oversold, and what assurance a regulated care setting has to demand before relying on it.
Data governance19 May 2025UNICEF Innocenti
How governments can use sensitive education data to improve outcomes without centralising it, and what governance that actually requires.
Learning systems1 January 2019UNESCO MGIEP
On building a platform that connects knowledge across an institute's work, and what machine learning could and could not do for it.
I am glad to hear from researchers, institutions and policy teams working on the same problems.