The Case
Social media are one important source of data that can be used for big data analysis and AI techniques. While most of us use social media to stay in touch with friends and acquaintances or to sustain professional networks, it is also possible to find additional insights from such data that can be used in ways beyond those originally foreseen. In our case, a new start-up company utilised global social media data to develop insights into current and upcoming social events (e.g. strikes or labour unrest) that can affect global supply chains. Such information is potentially valuable for international logistics companies, but also for many other organisations that rely on global supply chains. The start-up includes non-English social media feeds in its analysis, thereby allowing the development of risk intelligence in areas that are often difficult to access by Western companies that do not have proficiency in local languages outside of Europe and that do not have a strong local presence.
Ethical Issues
The start-up realised from the outset that its work would raise ethical issues. Key among issues raised in the literature but also understood by the company are:
- Privacy and data protection: Social media data, even when publicly available, can contain sensitive personal data that can be misused or have negative consequences for the data
- Transparency, accountability and bias: The prediction of social events needs to be supported by reliable data and done in a way that inspires confidence in This means that possible biases need to be recognised and addressed and that users need to be able to understand the approach and scrutinise it to some degree.
- Intervention effects: The prediction of events caused by often difficult labour relationships have the potential to influence these relationships and, for example, weaken or strengthen the role of employers or trade
Mitigation Strategies
The start-up understood early on that successfully dealing with ethical challenges is key to its success. It therefore adopted an approach to responsible data science that relies on three pillars:
- Establishment of an ethics committee comprising external and independent experts to scrutinise their approach;
- Development of a code of ethics, supported by the ethics committee, which forms part of contracts with clients, thus ensuring that ethical considerations are reflected across its business activities;
- Institution of stakeholder dialogue that involves relevant stakeholders across the various ethical issues of interest. These include employer and labour representatives to ensure labour relationship issues are clearly visible.
Lessons Learned
The case clearly demonstrates that ethical issues can go to the heart of big-data and AI-enabled businesses. However, an awareness of such issues and the ability and willingness to address them is not confined to large and international companies with extensive legal or CSR departments. Where awareness and willingness is present, they can be successfully addressed in any context, including that of small start-up companies. Specific insights arising from this case study include:
Proactive engagement with ethical issues not only helps avoid risks for a company, but can also offer business The start-up at the heart of this case study is now working with labour unions to develop new services for this type of user. Ethics can thus point to new business models.
Whilst ethical issues such as privacy apply to almost all AI cases, what counts is how general awareness is enhanced with context-specific
Existing mechanisms to reach out to stakeholders and draw on external expertise, which were present in the start-up environment of this company, can help organisations navigate the complex ethical issues of big data and
