Since late 2025, the US and EU have been working on an agreement to start sharing people’s personal data for border and migration purposes.
This ‘framework agreement’ is not intended to facilitate data exchange itself. Rather, it would lay the groundwork and standards for individual EU member states to negotiate their own deals with the US. Given its cross-EU nature, the European Commission was tasked with negotiating the agreement on behalf of the EU overall. The final document will serve as a baseline, but will not prevent EU member states and the US from exchanging additional information via other agreements.
In May 2026, Statewatch published the first draft proposal for this framework agreement. The analysis made it clear that the proposal went against EU law. Specific concerns included how the proposal could violate fundamental rights protections and access to effective remedy.
For example, the first draft appeared to give the US a lot of freedom to use EU data to profile or discriminate against EU citizens in ways they would not expect from their own governments. This could enable the US to apply racist profiling during travel or visa processes or to target peaceful activists.
On top of that, the first draft appeared to include terms that the Commission was not legally allowed to offer the US when it was tasked with the negotiations.
Minimal changes to a bad deal
After further negotiations, the US and the European Commission produced a second draft data exchange agreement in June.
This new draft—leaked to Statewatch—does not meaningfully address any of the issues in the first draft proposal. It shows only a few minor changes.
First, the purposes for which data can be exchanged have been limited to “serious and genuine risk to public security and public order”. This is an improvement over the more general reference to “genuine risk” in the first version of the agreement, which could have been interpreted broadly by over-eager authorities.
Nonetheless, the concepts of public security and public order are vague and could include legitimate expressions of political opinion. There is, for instance, serious concern that people might be targeted in the US for taking part in protests in Europe. Imagine pro-Palestine rallies, or indeed European protests against US government practices.
The second improvement is that the draft now includes references to “national information systems” under Article 5, instead of a general reference to information systems. This is likely an effort to exclude Europe-wide information systems from the agreement, something that would violate the negotiating mandate given by the Council of the EU to the Commission.
The revised agreement also introduces a very soft “minimum safeguard” for photographs. The new text specifies that a photo of somebody may be exchanged, in the context of identifying them via an automated search, but only if it is deemed ‘necessary’.
Finally, the revised agreement introduced only one change to the protection of personal information: a yearly review of information retention. The previous agreement left the review period to a more general term, “periodic.”
US ‘strongly insists’ on deregulating EU standards
The contextual information about the negotiations added to the revised agreement indicates that the US “strongly insists on not introducing additional limitations” to the exchange of data. In other words, the US want as much EU data as they can get.
Indeed, as Statewatch explained with the first version of the agreement in May, the EU and US approaches to data protection differ significantly. In July, Statewatch joined in pressing the Council of the EU to step in and not let the Commission cede to US pressure via an open letter coordinated by European Digital Rights (EDRi). Together, we warned:
“In its current state the draft Framework Agreement risks not standing up in court, as it conflicts on many counts with EU data protection legal criteria, as interpreted by the (Court of Justice of the EU).”
As revealed in the second draft agreement, the US has been creative in trying to limit protections for sensitive information. This includes data on people’s race or ethnicity, sexual orientation or political stances or speech. The US negotiators argue there should be no prohibition of the transfer of such information because “such information may indirectly be revealed as part of the information exchanged and should therefore not be categorically excluded.”
This statement is revealing in itself. It aligns with the critical literature on automated technologies by explaining how datasets can inadvertently reveal sensitive information. For example, certain postcodes can indicate a person’s race based on the demographics of that area. This observation has led many critics to argue for an even more careful and controlled system of government data collection.
In contrast, the observation seems to have led the US to feel justified in giving up on protecting people’s data altogether.
By offering no counterarguments to this perverse logic, the European Commission appears willing to acquiesce to it. In its rush to get a deal done, it seems to want to allow any data the US wants to be exchanged freely, with no further protection. Aside from the obvious privacy implications, this places the Commission’s stance in direct contradiction with EU law, which generally prohibits the use of citizens’ data in the way the US wants.
