From ethical data to action

5 ethical risks of using data and a concrete solution

  • Article
  • Data Science
Marieke Schulte
Chair Digital Power Datahub
6 min
16 Jul 2019

The introduction of the new privacy law (GDPR) in 2018 has ensured that many organisations put privacy high on the agenda. In this article you can read about the 5 ethical risks of working digitally and using data. We also share a concrete solution: the Responsible Data Framework.

The introduction of the General Data Protection Regulation has made many organisations rethink the way they use data. Violating the GDPR does not only result in large fines. It also poses increasing risks to the reputation of organisations. This is not to say that organisations now also deal responsibly with data out of intrinsic motivation. Not everything that is legal within the framework of the GDPR is also ethically responsible.

Luciano Floridi (2019)1, Professor of Philosophy and Ethics of Information and Director of the Digital Ethics Lab at the University of Oxford, describes in a recent study 5 ethical risks of working digitally and using data. 

Below we briefly go through these risks and suitable solutions with you. We then share a concrete solution for ethical risks of digital work with our client RNW Media, a Dutch social organisation. Here a Responsible Data Framework is used.

1. Ethics Shopping 

The first ethical risk associated with digital and data-related work is, according to Floridi, a consequence of the large amount of ethical principles, codes and manuals surrounding data. This proliferation of documents creates confusion among stakeholders. The risk of this is that organisations will go shopping and choose what suits them best. This allows them to mix and match ethical principles until they fit the work they are already doing. They do this instead of reviewing their work to conform to an ambitious set of principles.

Possible solution: Clear and large-scale ethical standards, such as the 'Ethics Guidelines for Trustworthy AI' recently published by the EU. 

2. Ethics Bluewashing 

As a variant of the term greenwashing, Floridi refers to the thickening of the ethical practices of organisations as ethics bluewashing.The implementation of ethical policies around digital work is often expensive and the returns are uncertain. That is why some organisations mainly focus on smart marketing. They then mask certain practices and advertise with fun ethical activities. Think of setting up an advisory body (which subsequently receives few powers). Although this looks good at first glance, it hardly adds anything in the end.

Possible solution: Transparency and education, for example in the form of campaigns on ethical practices. 

3. Ethics Lobbying 

Government policy often lags behind the latest developments in the digital world (due to asymmetric information). It seems that organisations are trying to lobby against the introduction of new, stricter legislation. They do this, for example, by introducing their own ethical policy. They then present this to policymakers as if government policy is no longer necessary. While this is a strategy that works mostly in the short term, it can cause a lot of damage. When government policy on digital work is delayed by lobbying, organisations are given free rein to develop their own policy and whether or not to comply with it.

Possible solution: Appropriate legislation and strict implementation thereof. Also: making ethics lobbying public when it takes place. This makes it easier to distinguish lobbying from genuine internal policy at organisations.

ethische data

4. Ethics Dumping 

Some organisations circumvent ethical reasons by moving their research activities on digital processes, products or services to countries where the relevant legislation is even less strict (often outside the EU). As a result, they are not inhibited by policy. They can also apply the results of their research later on in countries where legislation is strict. An example of this: a company trains new facial recognition algorithms on personal data of non-European citizens in a non-European country. It then applies these algorithms within the EU. This second step, importing unethical products or services, is especially difficult to control.

Possible solution: Governments can prevent irresponsible research from receiving subsidies through stricter controls. Quality marks for ethically responsible products or processes can prevent the import of irresponsible research.

5. Ethics Shirking 

Finally, Floridi describes the risk of ethics shirking.By this he refers to the selective application of ethical policy regarding digital work in areas, or in areas of work, where it pays the most. As with ethics dumping, vulnerable populations in developing countries are particularly affected by this. Here, institutions and organisations are not yet sufficiently equipped to protect the population against the risks of the digital world. That is why companies here invest the least in ethically responsible working.

Possible solution: Clarifying and naming responsibilities. Because responsibilities within organisations are unclear, processes can be outsourced more and more. This needs to be addressed to stop Ethics Shirking.

A concrete solution: the Responsible Data Framework of RNW Media 

A concrete solution is the Responsible Data Framework which we have been working on for our client RNW Media. RNW Media is a social organisation that mainly focuses on digital solutions for young people, by building online communities for specific themes, such as sexual health and rights. 

The themes for which RNW Media builds social media platforms are often sensitive in the countries where they are used.

An example of a risk associated with this is that a government with an anti-LGBTI agenda will obtain a database containing the personal data of LGBTI young people who use a platform from RNW Media. RNW Media therefore pays a lot of attention to the protection and responsible processing of data of those using its platforms.

We developed the Responsible Data Framework (RDF) as a basis for internal policy in which responsible handling of data is laid down. RNW Media's RDF is based on the files of other civil society organisations, such as Oxfam.

This makes it easy to compare RNW Media's policy with that of other NGOs, while still focussing on the organisation itself. RNW Media's Responsible Data Framework consists of 6 ambitious principles:

  1. Respect for data subject's autonomy 
  2. Respect for local context 
  3. Do no harm 
  4. Privacy principles and FIPs 
  5. Transparency and accountability 
  6. International considerations 

These six principles have specific objectives. Policy is then linked to them. An example is this objective under principle:

2. Respecting local context: 

  • We respect cultural norms and power differentials:
  • Individuals are agents in the data collection process; and
  • We are aware and considerate of the power differences at play between the data collector and the data subject.

From internal policy to sector-wide declaration for ethical data 

Colleagues from the Data & Digital team at RNW Media are part of a team of experts in AI & Machine Learning. There are also experts in social work, who recently launched The Dagstuhl Declaration on the Application of Machine Learning and Artificial Intelligence for Social Good

This statement bridges the gap between Data Science and the Sustainable Development Goals, and the people who work in both sectors. It is therefore a big step in the right direction.

RNW Media's Responsible Data Framework was one of the sources of inspiration for the statement. These kinds of initiatives are important for raising awareness of the ethical risks of digital working. They are a requirement for developing appropriate solutions for ethical data.

RNW Media is well on its way to not only preventing the ethical risks of working with data internally, but also to promoting a responsible data climate together with other organisations.

Have you been inspired by this article and would you like to discuss ethical data with one of our colleagues? Contact us.

1 Floridi, L (2019). Translating Principles into Practices of Digital Ethics: Five Risks of Being Unethical. Philosophy & Technology 32:185-193.

This is an article by Marieke Schulte, Data Science Consultant, Digital Power

Marieke has a background in social and development policy and evaluating it, using (ethically sound) data.

Marieke Schulte

Chair Digital Power Datahub+31(0)6 10 93 54

Receive data insights, use cases and behind-the-scenes peeks once a month?

Sign up for our email list and stay 'up to data':