Allied Intelligence — 2026-08-20

Israeli Government Creates Fake Think Tank Hanover Institute to Systematically Poison AI Chatbot Responses on Palestine With 900K Dollar Contract and 100-Plus AI-Optimized Reports

BLUFState-sponsored content engineered to exploit LLM retrieval pipelines now costs under a million dollars per campaign, and detection depends on disclosure regimes most actors can circumvent.

Piro Inc., co-founded by "Inside Man" producer Daniel Rosenberg, created the Hanover Institute for Public Policy for the Israeli Government Advertising Agency, receiving $900,000 from Israel and publishing over 100 unbylined reports on Israel and Gaza since August 6 12. The reports use footnotes, data tables and a neutral tone that NewsGuard analyst Alice Lee said mimics a credible think tank built for AI chatbot citation, a practice Piro markets as "AI Story Optimization" 12. Responsible Statecraft ran 12 sampled Hanover Institute articles through the AI-detection tool GPTZero, which flagged 11 as AI-written with high confidence and one with moderate confidence 12. Politico reported, citing FARA filings with the Justice Department, that the campaign runs through Havas Media and Seattle AI-optimization startup Res, with RT citing a filed budget of $100,000 against Piro's separately disclosed $900,000 34. Politico's investigation found ChatGPT and Perplexity citing Hanover Institute material in Gaza-related queries; Rosenberg told Politico the work integrates accurate, sourced facts rather than targets AI systems 34.

Analysis
Piro Inc.'s Hanover Institute shows footnotes, data tables and neutral tone now function as an exploitable credibility proxy for large language model retrieval rather than a guarantee of independent research, confirmed by ChatGPT and Perplexity citing its material in Gaza queries. Producing over 100 reports for under a million dollars makes the technique replicable by any government or advocacy group with a modest public relations budget, and detection depended entirely on a FARA filing obligation that domestic cutouts and non-FARA jurisdictions do not face; AI platforms still lack an attribution layer flagging sponsored content at citation, leaving exposure to investigative journalism rather than the chatbot providers. The record rests on a single primary report, Politico's newsletter citing DOJ filings, with other outlets amplifying rather than independently verifying; disclosure of Piro's government sponsor and unrebutted underlying claims support a transparency rather than covert-manipulation reading.

5 sources

Israel creates fake think tank in likely attempt to dupe AI chatbots - Responsible Statecraft
Israel set up fake think tank to sway AI chatbots: Report - Daily Sabah
Israel wants ChatGPT to push its Gaza narrative – Politico - RT
Israel is trying to shape what AI chatbots say about Gaza and the IDF - CTech (Calcalist)
Israeli PR Teams Want To Answer Your ChatGPT Questions - Politico (Politico Influence newsletter)