An investigation by the Quincy Institute for Responsible Statecraft found that Israel's government paid a firm to create a fake think tank designed to feed AI chatbots favorable content. The website, called the Hanover Institute for Public Policy, was set up by Piro Inc., a firm co-founded by film producer Daniel Rosenberg, on behalf of the Israeli Government Advertising Agency, according to the investigation.
The Hanover Institute has published more than 100 unbylined articles since August 6, formatted with footnotes and tables of contents in a neutral tone designed to appeal to AI systems, Responsible Statecraft reported. A small disclaimer discloses the Israeli government's involvement, according to the investigation.
Politico first reported a Foreign Agents Registration Act filing showing Piro received $900,000 from the Israeli government for the work, Arab News reported. Testing by Politico found that both ChatGPT and Perplexity cited Hanover Institute material when asked questions about Gaza and antisemitism, according to Arab News.
AI detection service GPTZero flagged 11 of 12 randomly sampled Hanover articles as AI written with high confidence, Responsible Statecraft reported. Piro's own website advertises an "AI Story Optimization" service, according to the investigation. Rosenberg told Politico the work aims to "put accurate, sourced facts into the public record" and counter misinformation, Arab News reported.
For builders working on retrieval augmented systems or AI search, this is a live example of content engineered specifically to read as neutral, well sourced and citable to a language model. Any pipeline that treats footnotes and a neutral tone as a signal of credibility has exactly the blind spot this campaign was built to exploit.