The draft asks governments to admit that punishing false speech can hurt true speech, which in Brussels now counts as a diplomatic incident.
American diplomats handed around a piece of paper asking governments to stop jailing and fining people for saying things that turn out to be wrong. European lawmakers received this as a hostile act.
The draft declaration, which the Trump administration wants signed on the sidelines of the UN General Assembly in September, asks signatories to “commit” to “recognize that punishing allegedly false speech” like “disinformation,” “misinformation” or “hate speech” can endanger “freedom of expression,” with a carve-out for speech that directly incites violence.
Under Secretary of State Sarah B. Rogers said her office circulated the text “in the margins of OECD” talks. Politico reported that the document also asks governments to back off demands that platforms monitor and delete what users post.
The operative sentence demands very little. It does not ask anyone to repeal a law but it asks governments to acknowledge that locking someone up for a false statement might have consequences for the ability to make true ones. That acknowledgment is now controversial enough in Brussels to count as a diplomatic provocation, which tells you something about where the argument has traveled.
The thing everyone understands the draft to be aimed at is the Digital Services Act, the EU law that requires the largest platforms to identify and reduce what the text calls systemic risks, a category that expressly includes disinformation and civic discourse.
Fines run to six percent of global annual revenue. The European Commission decides what counts as adequate mitigation. Nobody has to be convicted of anything, no court has to find that a specific post was false, and the pressure lands on the company rather than the speaker, which is the point of designing it that way.
A platform facing a nine-figure penalty for insufficient risk mitigation will delete first and litigate almost never. That produces the chilling effect without ever producing a defendant. The user whose post disappears gets no hearing because the user was never the target. The Commission never has to define disinformation with any precision, because it never has to prove a single instance of it in front of a judge.
The definition lives in guidance documents and compliance reviews, and the companies guess at where the line falls, and they guess conservatively, because guessing wrong costs six percent of everything.
Officials point out that the DSA passed with overwhelming support in the European Parliament and unanimous backing from all 27 member states.
Both claims can be true at once. A law can be democratically enacted and still hand a regulator the power to determine which speech a platform must suppress. Elections don’t convert that power into something else. The question of who writes the definition of “disinformation” survives every vote, and in this case the answer is a Commission directorate in Brussels that no European voted for and cannot remove.
What the administration is offering is nonbinding. A UN declaration creates no cause of action, overrides no statute, and binds no regulator. Contrast that with the DSA, which is enforceable today, with penalties calculated as a share of global revenue, against companies that host most of the world’s political conversation. One of these documents can silence someone and the other cannot, and the one that cannot is the one generating the outrage.
Whatever the motives behind circulating it, and there are obvious commercial ones involving American companies that would prefer smaller compliance bills, the sentence in the draft is correct. Punishing allegedly false speech does endanger the ability to speak. Governments that reserve the authority to declare statements false and impose costs for making them will use that authority, and they will use it most often against the people with the least ability to absorb the cost.