How Israel is trying to shape what AI tells users about Gaza
Israel has launched a new campaign to influence how large-language models like ChatGPT are answering questions about Gaza and the Israel Defense Forces amid declining public support for the country across the U.S. political spectrum.
— French PR firm Havas Media, which runs the lion’s share of Israel’s Foreign Agents Registration Act-registered foreign influence work in the U.S., appears to have stood up an “institute” aimed at feeding LLMs positive information about Israel via one of its subcontractors, boutique ad agency Piro, Inc.
— In a trove of FARA documents filed with the Justice Department in the last week — which have not previously been reported — Piro disclosed that over a dozen articles on the Hanover Institute for Public Policy’s website were part of a $100,000 campaign focused on “the creation and dissemination of factual, source-supported informational materials intended to educate the U.S. public regarding Israel and related issues through publicly distributed content.”
— The institute’s website features dozens of “data reports” with no named authors, each of which features questions like “Is the IDF the World’s Most Moral Army?” and “Is There a Policy of Starvation in Gaza?” — a telltale sign that the articles are designed to feed data to LLMs. The site’s “About” page says “The Hanover Institute for Public Policy studies the inputs fueling antisemitism in the United States, and publishes what the evidence shows,” adding that its unidentified founder “traces this work to a moment in childhood.”
— It appears to be working: ChatGPT and Perplexity both cited the Hanover Institute’s material in response to neutral tests run by PI, referring to the site’s content on Gaza, anti-Zionism and antisemitism. [Continue reading…]