WHEN SCIENCE GAINS ACCESS, WHO DECIDES WHAT COUNTS AS DATA?

Nov. 2026

A researcher wants to know how a digital platform shapes what millions of people encounter. The company offers an interface, an application form, perhaps a dataset. But the researcher must first determine which questions can be answered using information the company has decided to make available.

An October 2026 investigation surveyed 119 researchers worldwide and interviewed 19 in depth. Participants reported complex application procedures, difficulties obtaining credentials, and restricted data access. These barriers also reinforced inequalities between research institutions, countries, and available resources (Mimizuka et al. 2026).

The European Union's Digital Services Act has established a legal route for qualified researchers to request platform data. This represents genuine progress. Yet another 2026 investigation identifies a deeper obstacle: companies often define data through predefined, aggregated datasets, while independent researchers may need information about how the systems themselves operate (Annabell 2026).

The contradiction extends beyond corporate secrecy. An institution can become formally more transparent while retaining substantial influence over what transparency actually reveals.

Researchers may receive answers, but only to questions compatible with the available records. The operations they cannot observe may remain outside scientific investigation altogether.

Neither study demonstrates that every restriction is unjustified or that European regulation has failed. Privacy and security safeguards remain essential. But the evidence exposes a problem at the heart of Dark Machinic Epistemology: access to selected information does not necessarily provide access to the mechanisms producing it.

A corporation may disclose increasingly detailed information about its users while leaving the organization of its own interventions comparatively inaccessible.

What happens to scientific independence when the object of investigation also determines which parts of itself can be investigated?

Liviu Poenaru

References

Mimizuka, Kayo, et al. 2026. “Post-Post-API Age: Studying Digital Platforms in Scant Data Access Times.” Proceedings of the ACM on Human-Computer Interaction 10 (6): CSCW137. https://doi.org/10.1145/3816985.

Annabell, Taylor. 2026. “Regulating a Floating Signifier: Interpretations of ‘Data’ in the Digital Services Act Researcher Access Provision.” Media, Technology and Governance. https://doi.org/10.1177/29777984261486601.