Pierri, F., Araujo, T., Kruikemeier, S., Lorenz-Spreen, P., Abeele, M. M. P. V., Vandenbosch, L., Gonçalves-Sa, J., & Grabowicz, P. A. (2025). Research opportunities and challenges of the EU’s Digital Services Act. arXiv [cs.CY].
Summary
This commentary examines the practical implementation of Article 40 of the EU’s Digital Services Act (DSA), which grants vetted researchers access to data from very large online platforms (VLOPs) and search engines (VLOSEs). Written by researchers directly embedded in EU policy processes, the piece argues that while Article 40 is a landmark mechanism for platform transparency and accountability, substantial legal, technical, and organizational barriers threaten its effectiveness. The authors frame the DSA as a distinctively European model of platform governance, contrasting with US-based deregulatory trends, and contend that its success is essential for democratic oversight of digital infrastructures. They organize their analysis around three categories of challenges — misaligned incentives, resource asymmetries, and implementation bottlenecks — and offer concrete recommendations to address them.
Key Contributions
- Provides an informed practitioner perspective on the operational state of DSA Article 40 implementation as of late 2025.
- Structures the policy debate around three categories of challenges: misaligned incentives, resource asymmetries, and implementation bottlenecks.
- Offers three concrete recommendations: streamlining access with standardized data-sharing frameworks; supporting independent research through dedicated funding and community coordination; and extending DSA scope to large language models.
- Frames LLMs as intermediary services that should fall within DSA oversight, identifying a specific regulatory blind spot.
Methods
The paper is a collective expert commentary synthesizing the authors’ direct participation in EU policy processes. It draws on an 18-month ERC-supported pilot program connecting researchers, DG-CONNECT, and national Digital Services Coordinators (DSCs), alongside roundtable discussions with major platforms and the European Commission. These experiential insights are combined with a review of literature on platform research, industry influence in AI, and DSA legal scholarship.
Findings
- Platforms have shown varying reluctance toward direct data sharing, sometimes alluding to potential legal action against researchers and NGOs.
- A documented case (Meta’s collaboration with researchers around the 2020 US election) shows how undisclosed algorithmic changes can bias platform-mediated research conclusions.
- National DSCs and local university Data Protection Officers face likely mismatches in language and expectations with researchers and platforms, slowing first data requests.
- Industry AI research has overtaken public research in financial and infrastructural resources, while social media campaigns framing the DSA as censorship have reached tens of millions of users.
- A circular resource problem disadvantages early-career and under-resourced researchers: data access is needed to win grants, but funding is needed to pursue data access.
- The first decisions on data access requests via the DSA portal are expected in late February 2026.
Connections
This commentary is a keystone reference for the platform-governance-data-access literature, directly informing empirical work that depends on or evaluates platform data access regimes such as Bruns2026-yv and Rieder2026-pp. Its warning about “independence by permission” and biased platform-mediated research echoes concerns central to studies of independent research access and platform transparency. The paper’s framing of large language models as a systemic-risk regulatory blind spot connects it to the broader information-disorder conversation about emerging AI-driven manipulation risks.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts