Pro-Israel website floods web with reports aimed at influencing AI chatbot responses
Fake think tank published more than 560,000 words in nine days as part of an effort to shape how AI systems answer questions about Israel and Gaza
LONDON (MNTV) — A pro-Israel website operating under the name of a purported think tank has published hundreds of reports in a short period in an apparent effort to influence how artificial intelligence (AI) chatbots respond to questions about Israel and the war in Gaza, according to a new analysis.
The website, which presents itself as the “Hanover Institute for Public Policy,” published 124 reports containing more than 560,000 words between Aug. 6 and 14, according to an analysis published Wednesday by The Guardian.
The publishing surge was particularly intense on Aug. 12 and 13, when the website posted 73 reports totaling nearly 354,000 words.
The website addresses contentious issues including the torture of Palestinian prisoners, Israeli war crimes and whether Israel has deliberately starved Palestinians in Gaza, while presenting its positions as neutral academic research.
However, the purported institute has no physical address, named staff or bylines.
The Guardian also found no evidence that it exists as a legally registered entity in any jurisdiction.
The website is reportedly part of a broader Israeli government-funded campaign designed to influence how AI systems process and present information about Israel. Tens of millions of dollars have reportedly been channelled through third parties, including advertising company Havas Media.
Many of the website’s headlines are phrased as questions that resemble prompts users might enter into AI chatbots, including “Is Anti-Zionism Antisemitism?” and “Is Israel Carrying Out a Genocide in Gaza?”
The reports adopt an academic tone and cite institutions including UN agencies, the World Bank and Amnesty International, a strategy apparently intended to make the material appear credible and increase its visibility to AI systems.
The Guardian found, however, that the reports repeatedly frame information in ways that minimise criticism of Israel. References to allegations of genocide and apartheid, for example, are frequently followed by reminders that no final court ruling has established those claims.
Nick Cleveland-Stout, a research associate at the Washington-based Quincy Institute, said the website initially appeared highly professional but that closer examination revealed significant problems with its content.
Cleveland-Stout also warned that the campaign’s longer-term objective could be to place pro-Israel narratives into datasets used to train AI systems.
He said the effort could become more influential if the website’s material is indexed by repositories such as Common Crawl, which can feed data into AI training datasets.
The impact could extend beyond conventional online searches, he said, because AI systems may reproduce information absorbed during training without necessarily identifying where it came from.
Cleveland-Stout said the institute is likely to have some success in influencing chatbot responses but questioned whether such efforts would affect broader public opinion, which remains a separate challenge for Israel amid growing international criticism of its conduct in Gaza.