EU draft plan lets AI giants tap decades of citizen data

If this draft survives into law, it would be one of the most consequential pro-AI moves Europe has made, despite being framed as a simplification of digital rules. Treating AI model development and operation as a “legitimate interest” under GDPR would hand incumbent platforms the right to mine almost everything they have ever collected about Europeans, plus many data flows they can infer indirectly, with far less friction. For companies like OpenAI, Anthropic, Google and Meta, that is an invitation to scale training data and behavioral telemetry without constant consent pop-ups or legal contortions.

From the perspective of the race to AGI, easier access to high quality, long-horizon human data is rocket fuel. The biggest labs are already data-constrained at the margins. Being able to fold in years of messaging logs, social graphs, device identifiers and clickstreams within a broad “legitimate interest” frame would sharpen models’ understanding of human behavior, language and edge cases. It would also tilt the playing field even further toward firms that already sit on gigantic data reservoirs.

The flip side is political risk. Activists like Max Schrems are already calling this a “digital expropriation” of Europeans, and any scandal around AI misuse of personal data could provoke a backlash that locks down future data sharing more harshly. For now, though, the draft underscores that parts of the EU establishment are willing to trade privacy leverage for perceived competitiveness in the AI race.