EU Proposal Would Expand Europol’s Data Access and AI Policing Powers

Summary

A European Commission regulation proposed on 24 June 2026 would give Europol broader access to personal data linked to current and potential cross-border investigations and make it a more operational police agency. The proposal treats safeguards such as limiting data to people and investigations demonstrably relevant to crime as administrative burdens, and would let Europol collect and filter data later rather than only store limited information in existing investigations. Statewatch reports that former officials have described parallel “shadow systems” that operated outside Europol’s main IT environment, while a Front-LEX complaint asks the European Data Protection Supervisor to stop them. The plans would also make Europol a hub for private companies developing digital policing tools, including technology for subverting encryption and automated profiling, and would allow it to apply directly for research funding such as Horizon Europe, whose 2028-2034 budget is stated as €175 billion. The Commission says expanded datasets and machine learning are needed to modernise law enforcement, but Europol has already received exemptions from parts of the AI Act’s monitoring, oversight and transparency requirements for high-risk AI. An EDPS opinion describes automated scripts that select and modify internet data, with some changes potentially left untracked, raising concerns under a Court of Justice ruling requiring automation in criminal investigations to be explainable and identifiable. The proposal also envisions a Police Shared Data Space, automated data loaders, broader access to Prüm II biometric and police databases, and systematic information exchanges with agencies such as Frontex. Statewatch links these measures to risks of stigmatisation, discrimination and weaker redress, citing past bulk transfers of migrant data and concerns about automated policing experiments. It notes that private-sector AI testing in Europol’s Innovation Lab would take place amid wider evidence that experimental systems can fail badly, including a cited case in which Flock misread license plates in 71% of police alerts. The article argues that expanding opaque data systems and weakening safeguards could increase arrests while also increasing miscarriages of justice and dependence on black-box technologies.