Fake US thinktank set up and funded by Israel sought to game AI for propaganda
A pro-Israel messaging website posing as a thinktank called the "Hanover Institute for Public Policy" has published over half a million words in nine days, apparently designed to prime AI chatbots to reproduce pro-Israel framing on contested issues such as Gaza and alleged war crimes. The New York firm Piro Inc has registered the site with the US justice department as material distributed on behalf of the Israeli government, and it forms part of a wider campaign, backed by tens of millions of dollars and routed through firms including Havas Media, aimed at shaping how AI systems discuss Israel.
The Hanover Institute has no verifiable legal existence, physical address, named staff or bylined authors, and its funding page now admits an earlier claim of independence was true only "when it was written". Guardian analysis found it published 124 reports totalling over 560,000 words between 6 and 14 August, with 73 reports alone appearing on 12 and 13 August. Most headlines mimic chatbot-style questions, such as "Is Israel Carrying Out a Genocide in Gaza?", and the reports cite credible sources like the UN and Amnesty International while consistently downplaying accusations against Israel, a pattern researchers say only becomes apparent on closer reading.
- Fake "Hanover Institute" thinktank pushes pro-Israel content to influence AI chatbots
- Israeli government-funded campaign published 560,000+ words in nine days
- Site has no real staff, address or legal existence, Guardian analysis finds
New here? Start with this
Thinktanks are organisations that research public policy and publish analysis. Artificial intelligence chatbots draw on information found across the internet, so the material they encounter can affect how they answer questions about political disputes.
Israel is the state at the centre of this story, and Gaza is a Palestinian territory where Israel’s military campaign has prompted allegations of war crimes and genocide, which Israel rejects. The Hanover Institute for Public Policy presented itself as an independent US thinktank, but reporting found no verifiable organisation, address or named authors behind it.
The story raises questions about how governments and their contractors try to influence online information, and how that material may shape chatbot responses. It also matters because readers may not recognise advocacy presented in the style of independent research.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
All governments conduct international communications campaigns to advance their interests and perspectives. This organisation registered its Israeli funding with the US Department of Justice and presented substantive, source-documented arguments rather than misinformation. Ensuring artificial intelligence systems—which increasingly shape public understanding—encounter Israeli perspectives on contested issues is a logical extension of standard advocacy work that nations across the political spectrum undertake.
The case against
This operation deliberately masqueraded as an independent institution whilst being entirely government-funded, and specifically engineered content to manipulate AI systems rather than engaging in transparent public debate. The combination of institutional deception and systematic AI-targeting goes beyond legitimate advocacy to constitute coordinated information manipulation. Such operations exploit the opacity of how AI systems learn and make decisions, corrupting the information ecosystem in ways that deliberate but transparent advocacy does not.