Skip to content
Roy Saurabh

Research

What I study, and why it hangs together

Four areas. They share a question: what has to be true, technically and institutionally, before an AI system can be relied on where it affects people.

  1. 01

    AI assurance and accountability

    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?

    7 related outputs
    • AI assurance
    • Regulation
    • Reproducibility

  2. 02

    Privacy-preserving collaborative AI

    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?

    3 related outputs
    • Federated learning
    • Privacy
    • AI assurance

  3. 03

    Human-centred adaptive systems

    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?

    4 related outputs
    • HCI
    • Learning sciences
    • Mental health

  4. 04

    Data governance for public-interest systems

    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?

    3 related outputs
    • Data governance
    • Children's data
    • Education

Projects

Research projects

Author2025-

An assurance toolchain for deployed AI

A connected set of reproducible tools for deployment-conditioned risk analysis, executable regulatory reasoning, evidence adequacy scoring and adversarial stress testing.

AffectLog

Doctoral research2021-2024

Wearable data and clinician wellbeing

Doctoral research on whether wearable sensor data can support mental-health risk monitoring for frontline healthcare professionals, and where the method breaks down.

Université Paris Cité