Skip to main content



Reporting period: 2018-03-01 to 2018-08-31

Problem: Global supply chains are the dominant way to organize production in the world economy. One of the most pervasive problems in this context is the violation of human rights in developing and emerging nations that provide labour inputs; often with very low wages. During the past three decades, the global supply chains have been governed by guidelines and initiatives that seek to promote socially responsible business conduct through industry self-regulation and voluntary participation. In particular, brands formulate codes of conduct (ethical rules) and require from suppliers to obey them. They hire professional auditors or NGOs to conduct occasional on-site inspections to ensure compliance with these norms. When there is a deviation corrective action is required. Thus, brands represent their corporate interests, determine how workers’ interests are protected, and hire those who are supposed to act on behalf of them. Obviously, this firm-centred system of rights enforcement entails substantial risks of biased, false, or incomplete reporting on labour and other social issues.

Why it matters: The lack of accurate and comprehensive information imposes rising challenges to (multinational) corporations, institutional investors, and public procurement agencies because workers and communities in developing and emerging countries are more prepared to resist rights infringements. Also, consumers are more sensitised to unsustainable and unfair practices, which impacts their buying decisions. At the same time, a global regime of ethical principles and guidelines is emerging that supports and leverages bottom-up pressure from citizens and consumers. For example, the UN Guiding Principles on Business and Human Rights, the OECD Guidelines for Multinational Enterprises, and the Public Procurement Directive 2014/24/EU of the European Parliament and of the Council all emphasize the need for effective social risk management with a view to protecting workers and other stakeholders’ legitimate rights and claims.

Project objective: Globalworks Lund AB developed a data analysis system that can satisfy the demand for information that is qualitatively, and quantitatively meaningful for social risk analyses; i.e. identifying the extent and pervasiveness of material risks. Every day workers and other stakeholders in developing countries write about their grievances on social media in order to protest, reach out for help, seek advice, or share experiences. These voices provide rich and detailed accounts of labour and other human rights violations.

We developed and tested together with pilot customers the prototype of a platform, social@risk™, which collects these voices and transforms them into social risk assessments of sectors, regions, and individual suppliers. social@risk represents an intelligent, fully transparent system where experts and stakeholders of an issue inform the algorithms that facilitate data structuring and analysis.

The overall project objective is to transform our prototype into a scalable system that covers major risk countries, key export-processing sectors, and multiple social media platforms. We seek to expand and further develop our analytical capabilities in collaboration with a network of pilot clients and other stakeholders. Collaboration with clients is a straightforward approach to remain focused on the value proposition of our service.
During the feasibility study, we planned, executed, and finalised three assignments all with a view to establish long-term alliances. In particular, we conducted:

- A supply chain screening of about 40 original design manufacturers and final assembly manufacturers in the Chinese electronics industry. This assignment was developed in collaboration with Electronics Watch and partly financed by its members.
- On behalf of the Swedish National Coordinator for Sustainability in Public Procurement, we tested our system in the Thai poultry sector.
- We conducted an explorative study in the Vietnamese garment sector on behalf of Fair Wear Foundation

In order to implement these projects, we advanced three main elements of social@risk: Data collection, human-machine interaction for text analysis, and expert knowledge management.

Data collection: A precondition for our system to work is a stable and reliable flow of information from social media into our database. During the feasibility study, we developed a program for multi-thread scraping and enhanced our capability to adapt web crawlers for any type of web-based resource.

Human-machine interaction for text analysis: The core of our solution is a data analytical system that facilitates interactions between intelligent machines and human experts. The way how machines and humans can communicate in order to make sense of large text volumes depends on the information density. We created programs for different data environments that structure and analyse information based on unsupervised topical modelling, semi-supervised machine learning and dynamic text analysis. In addition, we designed interfaces that enable experts to structure, categorise, and analyse large amounts of data.

Expert knowledge management: Experts are a vital part of our approach to social risk analyses and integrating them into social@risk has been the main focus of this feasibility study. To this end, we developed work procedures that we gradually standardise and document with a handbook and video tutorials. We extended our expert network from China to Thailand, Cambodia, Myanmar, and Vietnam.

Besides the above named three studies, we were involved in a research collaboration with Hong Kong University and the Chinese University of Hong Kong. In this context, we collected social media data from vocational school students in China for the project “Learning to Labour: Social Media and Migrant Labour Protection in Mainland China.” The purpose of this project, from our point of view, is to improve our data collection technologies and enhance data analytical capabilities. We continue to support this project with AI-based analytics.
There are more than 100 rating agencies that provide environmental, social, and governance (ESG) data; among them large firms such as Thomson Reuters and Morgan Stanley Capital International (MSCI). Also, the number of ethical audits conducted by multinational enterprises increases due to stricter regulations and increasing consumer awareness.

Social risk data are high in demand but neither the auditing industry nor ESG data providers have been able to open a window onto the factory floors of global supply chains. The lack of data for accurate and systematic social risk management puts businesses, institutional investors, and public buyers increasingly at odds with a tightening network of ethical principles at the UN, OECD, and EU level.

Globalworks aim is to fill the rising market gap with social@risk. In the future, social@risk will use millions of posts from tens of thousands of factory sites as the basis for predicting social risks associated with a supplier, industry, and/or location. social@risk identifies, quantifies, and predicts the likely occurrence of social risk types. It is the first grievance-based risk management approach in the market. The ultimate purpose is to generate predictive insights about practices that violate human rights. social@risk adds a fundamentally new quality to social risk management that allows for pro-active social engagement and rights protection.
fiber optics workers
visualisation of social media posts
ship breakers