
Fake Israeli-backed thinktank used AI tactics to steer chatbots
A Guardian-reported investigation reveals a non-existent Hanover Institute for Public Policy funded by the Israeli government and routed through Piro Inc., Havas Media and Israel’s LaPam to publish hundreds of thousands of words in nine days. The material, designed to look like neutral research, is intended to influence AI chatbots by optimizing content for citations, raising concerns about training-data poisoning and the potential for bots to repeat biased narratives. The Hanover Institute appears to have no verifiable existence, and its disclosures point to foreign-agent registrations under FARA. While Israel disputes conducting influence operations, analysts warn such campaigns could shape AI outputs and public diplomacy in ways that are hard to fact-check, especially as US opinion shifts on Israel’s actions in Gaza.






