Research Supervisors: Dr. Xingjie Wei and Prof. Chee Yew Wong
Background and Rationale
Modern slavery affects an estimated 50 million people worldwide. While the UK's Modern Slavery Act 2015 requires large companies to publish annual disclosure statements, regulators lack the resources to assess compliance at scale.
Recent advances in large language models and annotated datasets (AIMS; Bora et al., 2025) make automated analysis possible. This study applies semantic analysis to UK corporate statements, leveraging the richer Australian annotation schema to move beyond binary compliance. We assess disclosure quality, examine variation by industry and firm size, and explore patterns across reporting criteria.
Research Questions
- To what extent do UK corporate modern slavery statements fulfil the mandatory reporting criteria stipulated by the Modern Slavery Act 2015?
- To what extent do compliant disclosures demonstrate substantive intent, and to what extent do they rely on vague or unsubstantiated language?
- How do compliance rates and disclosure quality vary across company size and industry?
- Does the quality of disclosure in one mandatory reporting criterion predict compliance behaviour in others, and are there systematic patterns of selective or compensatory reporting?
Methodology
Using the publicly available AIMS datasets (Bora et al., 2025), we apply semantic analysis to UK modern slavery statements to assess compliance with the Modern Slavery Act, evaluate disclosure quality, and examine how reporting varies by industry and firm size. We also explore relationships between different reporting criteria to identify patterns of selective or compensatory disclosure. All data processing and analysis will be implemented using KNIME Analytics Platform to ensure a reproducible, modular workflow.