Cambridge Academic Statement Example: Algorithmic fairness researcher to AI policy (Score 93)

Translating algorithmic fairness evidence into actionable policy is, I have come to believe, one of the most consequential methodological problems in contemporary technology governance. That conviction did not arrive abstractly. It emerged from a specific moment during my undergraduate research project, when I found that a fairness metric I had applied to a hiring-algorithm dataset produced statistically clean results that were, on any plausible reading of the affected population, substantively misleading. The gap between technical correctness and normative adequacy became the organising question of my subsequent work, and it is the question I wish to pursue rigorously at Cambridge. My BSc in Data Science, with a focus on fairness evaluation, gave me the quantitative foundations for that inquiry. In my final-year applied project I conducted a systematic comparison of demographic-parity, equalised-odds, and counterfactual-fairness criteria across three public procurement datasets, finding that criterion choice shifted the policy-relevant conclusion in every case. That finding shaped a short recommendation note I prepared for an internal departmental working paper, where I argued that regulators who specify a fairness metric without justifying its normative basis are, in effect, making a hidden value choice. The paper is currently under internal review. Alongside this, a summer placement with a technology governance advisory team required me to translate that technical analysis into a briefing note for a non-specialist planning discussion, an exercise that exposed precisely where the academic literature on algorithmic accountability stops short of practical operationalisation. The MPhil in Technology Governance addresses that gap at the level I need. The programme's integration of sociotechnical analysis, regulatory theory, and empirical research methods matches the triangulated approach my work has demanded but that my undergraduate curriculum could only partially supply. I am particularly drawn to the treatment of data governance and platform regulation, where questions of measurement validity intersect with institutional design, and to the programme's emphasis on producing policy-relevant research rather than purely technical outputs. My intention is to use the dissertation to examine how fairness-metric selection could be embedded in algorithmic impact assessment frameworks, building on the comparative methodology I developed at undergraduate level while engaging seriously with the legal and political-economy literature the programme will open. I am aware that moving from technical fairness analysis to governance scholarship requires me to strengthen my grounding in regulatory theory and qualitative policy research. I have begun that preparation through directed reading in administrative law and science-and-technology studies, and I expect the programme's research methods training to consolidate it. Cambridge is the right environment for this transition because the combination of interdisciplinary supervision, proximity to the Bennett Institute's governance research, and a cohort drawn from both technical and policy backgrounds will hold my work to standards I cannot replicate elsewhere.