Holli Sargeant
@hollisargeant.bsky.social
Research Fellow in Law & AI, St John's College, University of Cambridge
Focusing on legal frameworks for algorithmic decision-making & role of AI in justice systems.
Focusing on legal frameworks for algorithmic decision-making & role of AI in justice systems.
Our results show that fine-tuned transformer encoders (ModernBERT F1 93.3%) substantially outperform both rule-based (RegEx F1 35.4%) and generative models (GPT-4.1 F1 76.6%), underscoring the value of targeted, jurisdiction-specific training for complex legal tasks.
#LegalNLP #LegalAI
#LegalNLP #LegalAI
November 4, 2025 at 9:49 AM
"Classifying Hate: Legal and Ethical Evaluations of ML-Assisted Hate Crime Classification and Estimation in Sweden" (w Hannes Waldetoft & @mansmag.bsky.social) studies where models succeed and struggle when classifying hate crime motivations in police reports.
facctconference.org/static/docs/...
facctconference.org/static/docs/...
facctconference.org
June 23, 2025 at 5:03 AM
"Classifying Hate: Legal and Ethical Evaluations of ML-Assisted Hate Crime Classification and Estimation in Sweden" (w Hannes Waldetoft & @mansmag.bsky.social) studies where models succeed and struggle when classifying hate crime motivations in police reports.
facctconference.org/static/docs/...
facctconference.org/static/docs/...
"Formalising Anti-Discrimination Law in Automated Decision Systems" (w @mansmag.bsky.social) tackles shortcomings in algo fairness approaches when applied to anti-discrimination law, introducing a novel decision-theoretic framework to mitigate unlawful discrimination. arxiv.org/abs/2407.00400
Formalising Anti-Discrimination Law in Automated Decision Systems
Algorithmic discrimination is a critical concern as machine learning models are used in high-stakes decision-making in legally protected contexts. Although substantial research on algorithmic bias and...
arxiv.org
June 23, 2025 at 5:03 AM
"Formalising Anti-Discrimination Law in Automated Decision Systems" (w @mansmag.bsky.social) tackles shortcomings in algo fairness approaches when applied to anti-discrimination law, introducing a novel decision-theoretic framework to mitigate unlawful discrimination. arxiv.org/abs/2407.00400