WAR ON GAZA
4 min read
Fake US think tank funded by Israel sought to influence AI chatbots: the Guardian
A pro-Israel website published 124 reports totalling more than 560,000 words in nine days in an apparent effort to shape AI-generated responses on Israel, a Guardian report reveals.
Fake US think tank funded by Israel sought to influence AI chatbots: the Guardian
(FILE) The website is part of a broader campaign funded by tens of millions of dollars from Israel. / Reuters

A pro-Israel website operating under the name of a think tank that does not exist has published more than 560,000 words in nine days in an apparent effort to influence how AI chatbots present Israel-related issues, a report by the Guardian revealed.

The Hanover Institute for Public Policy, a name under which the site operates, presents its material as neutral research on issues including Palestinian prisoners, Israeli war crimes and Israel’s use of starvation against Palestinians in Gaza.

But the site has no apparent legal entity, physical address, named staff or bylines. Its terms of use refer only to “the state in which the Institute is established” without identifying one.

New York-based media company Piro Inc registered the material with the US Justice Department under the Foreign Agents Registration Act (FARA) as content distributed for the Israeli government, the Guardian reported.

The website is part of a broader campaign funded by tens of millions of dollars from Israel and routed through third parties including Havas Media, according to FARA filings and other records. A Havas contract with another US firm involved in the campaign calls for the “deployment of websites and content to deliver GPT framing results on GPT conversations”.

Content designed for altering AI response

Between August 6 and 14, the Hanover Institute published 124 reports totaling more than 560,000 words, according to a Guardian analysis. It published 73 reports, totalling nearly 354,000 words, on August 12 and 13 alone.

Almost all of the headlines are phrased as questions that could be entered into an AI chatbot, including “Is Anti-Zionism Antisemitism?” and “Is Israel Carrying Out a Genocide in Gaza?”

The reports use a quasi-academic format and cite sources including the World Bank, UN agencies, the Palestinian Central Bureau of Statistics, Amnesty International and the Genocide Convention.

Nick Cleveland-Stout, a research associate at the Quincy Institute who has investigated the site, said in comments to the Guardian that it appeared designed to look academic despite what he described as significant flaws in its content.

The site's reports consistently frame criticism of Israel with caveats. According to the Guardian's analysis, 20 of the 124 reports for example note that no court has entered a verdict on claims that Israel is committing genocide or apartheid.

Another phrase, referring to casualty and aid figures as coming from “a party to the events”, appears in 55 reports and is applied to Israeli figures as well as Palestinian ones.

A 2022 study of antisemitic comments on European media outlets' Facebook pages is cited 454 times across 88 reports, including reports on subjects ranging from land registration and AIPAC to Gaza's death toll and the blockade.

RelatedTRT World - Israel funded $46.5M campaign to shape AI responses on Gaza war: report

Targeting to poison AI systems

According to the Guardian, the site initially used an AI-focused content platform called Res, which advertises technology designed to help websites get cited by ChatGPT, Perplexity, Claude and Gemini.

Piro also advertises an AI Story Optimization service that produces content “engineered for how LLMs evaluate credibility”.

Cleveland-Stout told the Guardian that the campaign could seek not only to make chatbots cite the websites but also to get the material into repositories such as Common Crawl, which can provide data used to train commercial AI models.

The Hanover Institute does not yet appear in Common Crawl's July index because its reports were published after the crawl closed, according to the Guardian.

Funding behind the campaign

Piro has a $1 million contract connected to the campaign, according to FARA filings.

Havas Media Germany has also been named as an intermediary for several US firms working on the Israeli account, including Clock Tower X, Targeted Communications Global, Bridges Partners and Davis Media NY, according to the Guardian analysis.

Justice Department filings show Havas paid Clock Tower X $15 million between October 2025 and March 2026, and Targeted Communications Global nearly $9.9 million between October and December 2025.

Israeli procurement records show Havas entities previously received tens of millions of dollars in contracts from LaPam, Israel's government advertising agency, for international media planning and purchases.

The wider campaign has also included social media efforts such as Freedom Not Terror, whose promotional video had received 28.1 million views on YouTube, the Guardian reported.

RelatedTRT World - Israel killed over 21,500 Palestinian children, destroyed 1,077 Gaza mosques: Media Office

Questions over effectiveness

The campaign comes as Israeli officials have expressed concerns about the effectiveness of efforts to improve Israel's image in the US.

Gallup found in February that Americans sympathised more with Palestinians than Israelis for the first time, while a Pew survey in April found 60 percent of Americans held an unfavourable view of Israel.

The Israeli Foreign Ministry has denied conducting influence operations in the US and said it follows US law.

Piro co-founder Daniel Rosenberg, responding to the Guardian’s questions, claimed the firm was retained to put “accurate, sourced facts” into the public record and counter “misinformation about Israel”.

SOURCE:The Guardian