18 november 2024
Executive Summary
On 4 November 2024, the Meijers Committee hosted an online conference titled “Ethnic Data Registration in the EU” to discuss safe and effective methods of ethnic data collection. The conference brought together academics, government officials, and representatives from Equinet, the European Network against Racism, and the European Data Protection Supervisor.
Keynote Highlights
Wojciech Wiewiórowski emphasized the ethical and legal complexities of ethnic data under General Data Protection Regulation (GDPR), stressing “do no harm” principle, voluntary participation, and the right to withdraw consent. Additionally, he noted the potential of the Artificial Intelligence Act (AIA), especially in mandating sensitive data collection for bias testing.
Linnet Taylor addressed the risks of ethnic data collection, particularly in state-controlled contexts where biases may lead to unfair treatment. She explained the need for voluntary participation, transparency, and community involvement in data collection, and recommended a “social license” for statistical offices to build public trust.
Panel Discussion
The panel, led by Zuiderveen Borgesius, covered key themes in ethnic data collection, including the need for reliable data, community engagement, and data protection under GDPR. They highlighted Europe-wide challenges, such as limited resources and the lack of political action on race equality. A call was made for unified EU guidelines on data handling to ensure consistency and protection, drawing on the AI Act for safeguards.
Key Points
The panel also discussed issues around collecting ethnic data in law enforcement, particularly the risks of using perception over self-identification, which can lead to misidentification. Belgium’s model of police training on “professional profiling” was cited as an approach to reduce racial bias.
Conclusion
The panel broadly agreed on the necessity of ethnic data collection, with strict safeguards, voluntary participation, and transparency. Unified EU guidelines are required to ensure ethical data practices, drawing on safeguards outlined in the AI Act. Concluding the event, de Vries highlighted that the core concern should be addressing structural injustices through purposeful data collection, ensuring data serves as a tool for meaningful policy improvement.
Introduction
On 4 November 2024, the Meijers Committee hosted an online conference, “Ethnic Data Registration in the EU,” to explore methods of gathering equality data with minimal risks. In academic and policy discussions, views on registering and collecting ethnic data vary greatly: supporters see it as crucial for tackling inequality, while opponents fear it could lead to discrimination and privacy issues. The conference focused on safeguarding individuals if such data collection proceeds, addressing key points from keynote speeches and panel discussions. Meijers Committee members Karin de Vries (chair of the day) and Frederik Zuiderveen Borgesius (chair of the panel) moderated the conference.
In her introduction, Karin de Vries, mentioned the EU Anti-racism Action Plan (2020-2025), which encourages Member States to gather disaggregated racial and ethnic data to help policymakers and the public understand and address discrimination. Such data, as emphasized by UN bodies and legal experts, can be crucial in supporting discrimination claims and aiding strategic litigation. However, significant concerns remain around aligning ethnic data registration (EDR) with privacy protections and the potential of perpetuating racial biases. Efforts so far suggest that data protection should not prevent EDR but that careful, community-inclusive approaches are essential.
Thanks to the speakers and panelists
We would like to thank all keynote speakers and panelists for their contributions: Wojciech Wiewiórowski (European Data Protection Supervisor), Linnet Taylor (Prof. International Data Governance at Tilburg University), Hielke Hijmans (Prof. at Vrije Univeristeit Brussel, President of the Litigation Chamber and Member Executive Board of the Belgian Data Protection Authority), Lina Mechbal (European Network against Racism), Daris Lewis (Legal and Policy Officer, Equinet) and Lilla Farkas (European Network of legal experts in gender equality and non-discrimination).
Keynote Speeches
Wojciech Wiewiórowski
Wojciech Wiewiórowski addressed the ethical and legal complexities of collecting ethnic and racial data, highlighting the EU’s General Data Protection Regulation (GDPR), which generally prohibits processing such data unless justified by public interest, research, or anti-discrimination efforts. The Regulation also includes strict safeguards ensuring adequate protection for individuals.
He emphasized the principles of “do no harm,” voluntary participation, and the right to withdraw consent, noting anonymization as a protective measure, albeit imperfect. He also flagged the paradox of self-identification, where personal identities may shift or contrast with external perceptions, potentially eroding trust in instances where such data is used by law enforcement.
Lastly, Wiewiórowski expressed concerns over the Artificial Intelligence Act (AIA) potentially mandating sensitive data collection for bias testing. He cautioned against normalizing such practices and underscored the need for careful, continuous dialogue.
Linnet Taylor
In her keynote, Linnet Taylor highlighted that while diversity and equality are prioritized, collecting equality data presents challenges. Different aims lead to varying degrees of disruption, including minimal institutional involvement, anti-discrimination enforcement, and proactive anti-racism efforts.
Taylor noted that ethnic data collection is particularly risky, as data accessible to state institutions may lead to unfair treatment due to structural biases. She proposed several principles improving data collection, namely voluntary participation, transparency, participant engagement, and options for updates and nuanced responses.
Further, she emphasized involving communities in decisions about ethnic data use, especially in policy. Effective data collection relies on trust, self-categorization, and care for respondents. She also recommended strengthening the “social license” for statistical offices through transparent community assessments, consent for secondary use, and veto rights for affected groups. Importantly, trust is often the hardest to achieve, given concerns over data misuse and the lack of clarity in regard to who manages the data and how those are handled.
Panel Discussion
The panel, featuring Lilla Farkas, Hielke Hijmans, Daris Lewis, and Lina Mechbal, led by Frederick Zuiderveen Borgesius, discussed key issues in ethnic data collection, including its importance, voluntariness, law enforcement’s role, and ways to reduce discrimination risks.
They noted challenges in collecting race equality data across Europe, especially the lack of political action, limited resources, and sampling difficulties faced by projects on ethnic minorities. While the Race Equality Directive addresses various social domains, it notably omits issues around descendants—an oversight impacting many European countries. Ongoing discussions with the Commission continue, yet this gap remains unaddressed.
The panel endorsed data protection as a safeguard for privacy and equality, particularly under GDPR, emphasizing the need for proportionate, purpose-specific use of sensitive data. Additionally, specific protections are crucial for data from vulnerable groups.
Anonymization and pseudonymization were recommended by Hijmans as a way to protect identities while preserving data utility, alongside adherence to clearly defined, legitimate purposes. Legal and technical solutions were also noted to balance data use with privacy, underscoring the need for strict safeguards and compliance.
Key points reiterated included the importance of community involvement in data collection and policymaking, transparency on data use, and addressing mistrust from vulnerable groups toward ethnic data collectors.
Panel coordinator, Frederick Zuiderveen Borgesius, emphasized that voluntary disclosure of ethnicity should not be problematic. Nonetheless, it was generally recognized that the topic of data collection, especially when related to ethnicity and race, is highly controversial. Indeed, many agreed that collecting such data can contribute to the stigmatization of certain groups.
The panel highlighted how a lack of complaints leads authorities to underestimate discrimination, with governments often relying on this absence to justify inaction. However, underreporting stems from low awareness of complaint mechanisms and mistrust in institutions and governments.
Daris Lewis stressed the importance of reliable data, citing FRA and INAR statistics on ongoing discrimination, including an 86% poverty risk for Roma children and high harassment rates for Jewish and Muslim women. Accurate data, he argued, is essential for effective anti-discrimination measures.
The panel advocated a fully inclusive approach to ethnic data collection, involving communities at every stage – from design to implementation. Lewis referenced a Eurobarometer survey showing 75% of EU citizens are open to sharing ethnic information, suggesting proactive engagement could enhance data quality.
The discussion continued by examining voluntariness, asking how it could apply in instances where individuals interact with law enforcement authorities.
Farkas discussed how analyzing existing records, rather than collecting data directly, could reveal patterns of racial profiling. However, it becomes more challenging when collecting data on people’s experiences of discrimination outside of law enforcement.
Lina Mechbal pointed out that there are already existing data sources on how racialized communities experience policing, such as surveys conducted by the Fundamental Rights Agency. These surveys, including ones focused on being Black or Muslim in the EU, have already included questions about police interactions. As such, she suggested that using existing surveys is a more appropriate and ethical way to gather information rather than relying on police to collect it during interactions.
Lewis highlighted the issues with collecting ethnic data in law enforcement, questioning whether data should rely on self-identification or perception. As such data often reflects authorities’ perceptions, there’s a risk of misidentification and misleading conclusions. He stressed the need for proper police training, citing Belgium’s approach where the equality body UNIA and Ghent collaborate to train officers in “professional profiling” to reduce racial bias. This partnership also monitors police adherence to these standards and refers to profiling complaints to UNIA. The strong need for standardizing data collection processes was highlighted since this can ensure transparency and provide a way for citizens to contest data related to racial profiling.
Wrap-up
Panel coordinator Fredrik Zuiderveen Borgesius, and Chairwoman Karin de Vries concluded the panel both of them summarizing key points and takeaways.
Overall, there was consensus on the need for ethnic data registration, though with careful attention to its risks. Many advocated for voluntary data collection, limited to specific purposes, with transparency and involvement of affected communities. Aggregated data can be useful if robust safeguards are in place. The EU requires unified guidelines for data collection, storage, and deletion, ensuring consistency and addressing known risks.
In closing, Karin de Vries emphasized that the primary concern should be the purpose behind data categorization, not the categorization itself. In her words “rather than focusing on whether or not we should engage in EDR, our starting point should be the injustices faced by minority communities. From there, we need to proceed by determining if, in which ways and with which safeguards, EDR can contribute to making the injustice visible and create room for remedies.”
