A website posing as a US think tank has published more than 500,000 words across 124 reports in just nine days in an apparent effort to influence how artificial intelligence chatbots present issues related to Israel, according to a report by The Guardian.
The website funded by Israel, operating under the name Hanover Institute for Public Policy, presents itself as a neutral research organisation. However, The Guardian found no evidence that the institute exists as a legal entity, while the site lists no physical address, staff or bylines.
Between August 6 and 14, the website published 124 reports totalling more than 560,000 words. On August 12 and 13 alone, it released 73 reports containing nearly 354,000 words.
ALSO READ: China-Linked Hackers Targeted NASA, Fed And Senate, Says US
Almost every headline is framed as a question that a user could ask an AI chatbot, including “Is Anti-Zionism Antisemitism?” and “Is Israel Carrying Out a Genocide in Gaza?” The reports use academic-style language and cite sources including UN agencies, the World Bank and Amnesty International.
However, researchers cited by The Guardian said the material consistently frames issues in ways that could reduce criticism of Israel. For example, reports frequently note that courts have not issued verdicts on allegations such as genocide or apartheid, while casualty and aid figures are often attributed to “a party to the events”.
The website initially used an llms.txt file associated with Res, an AI-focused content platform that markets services designed to help websites get cited by ChatGPT and other AI systems. The file was later replaced with a customised version, according to the report.
The website was filed with the US Justice Department under the Foreign Agents Registration Act (FARA) by New York-based Piro Inc, which disclosed work connected to the Israeli government. Piro is reportedly contracted for $1 million, while the broader campaign involves tens of millions of dollars routed through companies including Havas Media.
ALSO READ: Musk's X Suit Against New York Over Hate-Speech Data Dismissed
The campaign appears aimed not only at influencing chatbot search results but potentially at getting content into datasets used to train AI models.
The report comes amid growing concerns over attempts by governments, companies and other groups to shape AI-generated information by flooding the web with targeted content.
Israel has denied conducting influence operations in the US, while Piro has said its work is intended to put accurate, sourced information into the public record and counter misinformation about Israel.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.
