Two agencies attacked in five days, and the verb that keeps appearing is ‘says'

Four days before the Jalud attack, France24 and AFP reported “West Bank settlers attack AFP journalists.” The Times of Israel logged the same incident as “AFP says West Bank settlers attacked its journalists.” On 29 August the pattern repeated: the Times of Israel’s liveblog entry read “Footage shows settlers attacking Palestinian woman, NBC team in West Bank” — an evidentiary hedge with the video in hand — while Mediaite, a US media-trade publication, wrote that the NBC team “Reports Being Attacked.” Haaretz had already published the underlying dynamic on 25 August: “Settlers Attack Palestinians, Israeli Activists in West Bank as Soldiers Watch.” The Times of Israel reported the same week that settlers were “filmed attacking activists, reporters in front of IDF troops; none arrested.” Two international agencies struck in five days is what turns an anecdote into a pattern.

One artefact from the trade press is worth recording on its own. Journalism Pakistan, an industry outlet, published “Israeli settlers assault foreign journalists in West Bank” with an appended note stating the piece was based primarily on a WAFA report of 29 August, with context from Reuters reporting dated 29 August, an RSF statement of 20 August and a CPJ review published 14 November 2025 — and that its photograph is AI-generated, “for illustrative purposes only.” A journalism-industry outlet covered a filmed assault on journalists by illustrating it with a synthetic image. The same note observes that WAFA did not identify the journalists or give details on injuries or arrests, and that the Israeli military said troops intervened nearby but did not address the reported assault on the press.

Assessment: The hedge is not neutrality; it is risk allocation. “AFP says,” “NBC reports being attacked” and “footage shows” move the burden of the claim from the desk to the victim, and they do it at exactly the moment when video, named correspondents and an Israeli prime ministerial condemnation of the same incident have removed any factual reason to hedge. Distrust the reverse inference too: the presence of a claim verb does not prove bad faith, and house style at liveblogs routinely attributes everything. The AI-generated illustration is the sharper object. It is what happens when a story is aggregated rather than reported — and aggregation, not suppression, is the mechanism by which the agent in these headlines quietly disappears.