The goal of this project is to better understand and identify risks of modern slavery in global supply chains. On this page, we provide transparency on how the project, database, and software prototype were developed, as well as the limitations that should be considered when interpreting the results.
note: noch mit Bildern/Grafiken ergänzen
As modern slavery is a complex concept, organizations and scholars use different definitions to emphasis different forms of coercion and exploitation. To ensure conceptual consistency throughout the project, we developed a project-specific definition of modern slavery based on existing academic literature. This definition was refined and validated through two workshops with international experts on modern slavery.
Learn more about our definition.
To develop a comprehensive database of documented modern slavery cases, we systematically collected information from multiple public sources worldwide, including:
- News articles via nexis uni
- NGO reports via scarping sites of NGOs found on modern slavery map
- Court cases via sherlock
- Scientific papers (mostly case studies) in web of science and scopus
The cases were identified through a standardized search strategy that was iteratively refined throughout the project. Data collection focused on cases published between 01.01.2015 and 31.12.2024, available in English or German. The data was then filtered:
13,395 identified cases from 9,370 documents
-> 8,789 work-related cases (e.g., excluding forced marriage)
–> 2,193 sufficiently documented cases (e.g., excluding cases which were mentioned superficially)
—> 491 high-risk modern slavery cases included in the final database
For the last step, the cases were analysed by screening for high risk of modern slavery to exclude cases that do not meet the project definition. This again took various refinements. Finally, the high-risk cases were analysed along various categories, such as the country of exploitation. These categories were reviewed and validated by international experts on modern slavery.
To further improve reliability, experts independently evaluated cases and assessed whether they met the project definition of modern slavery. This validation process helped refine both the database and the classification of cases.
Artificial intelligence supported the identification and processing of cases. However, all key design decisions, filtering procedures, and validation steps involved extensive human review and iterative refinement. Furthermore, it is important to note that the database does not represent all instances of modern slavery worldwide. It is based on documented cases available in public sources and is therefore influenced by reporting practices, data availability, and regional differences in documentation.
See here information on possible biases during the process
The software prototype was developed using the validated case database as its foundation. The objective is to support users in identifying potential risks and relevant patterns while making uncertainties and limitations visible. The prototype combines: (topic modelling etc.), ethical AI workshop, )
The database does not contain all cases of modern slavery: Modern slavery is often hidden and underreported. The database is based on documented cases available through public sources and therefore reflects existing reporting and documentation practices. Cases that are never reported or documented cannot be captured.
Data availability differs across regions and sectors: Some industries, countries, and forms of exploitation are documented more frequently than others. As a result, some forms of modern slavery may be overrepresented while others remain underrepresented. Example?
AI supports, but does not replace, human judgement: The software prototype is intended as a decision-support tool. Its outputs should not be interpreted as definitive measurements. They should not be used without verification, as the database and the tool may contain errors. Human expertise remains essential when assessing risks and interpreting results, especially in the sensitive field of modern slavery.
others?