JD Vance said during nuclear negotiations that Iran submitted three different versions of a 10-point diplomatic proposal - and one of them appeared to have been written by ChatGPT.
The claim, made public by the US Vice President, adds an unusual detail to ongoing Iran nuclear talks. Sending multiple versions of the same proposal suggests a strategy of testing different framings on the other side. The ChatGPT version, according to Vance, was visibly distinct - presumably recognizable by the smooth, structured prose that AI-generated text tends to produce when compared against human-authored versions of the same document.
Using AI to draft official communications isn't surprising in 2026. ChatGPT and similar tools are used by teams across business, law, and government to draft, revise, and format documents. What's notable here is that the AI-generated version was apparently identifiable - either by stylistic signature, or because the writing shifted noticeably across the three versions in a way that made the origin obvious.
The diplomatic interpretation is harder to read with certainty. Submitting a ChatGPT-drafted proposal in formal nuclear negotiations could reflect internal time pressure, a disagreement within the Iranian delegation about what to put forward, or a deliberate signal that the proposal wasn't intended as a serious final position. Vance's decision to highlight it publicly suggests the US side read it as the latter.
For anyone who regularly uses AI-written content in professional contexts, the episode makes a practical point: AI-generated text is more detectable than most people assume, particularly when the reader has other writing from the same source to compare it against. A sudden shift to polished, structure-heavy prose reads differently than a document with the friction and idiosyncrasy of human authorship.