LONDON: A non-existent US think tank published more than half a million words of pro-Israel material in just nine days as part of an Israeli government-funded campaign designed to influence how artificial intelligence chatbots respond to questions about Israel and Gaza, The Guardian has found.
The Hanover Institute for Public Policy published 124 reports between Aug. 6 and 14, presenting them as neutral, academic-style research on contentious issues including allegations of genocide, war crimes, torture of Palestinian prisoners and the starvation of Palestinians in Gaza.
The analysis by The Guardian revealed that there was no evidence that the institute existed as a legal entity in any jurisdiction. Its website listed no physical address, staff members or authors for its reports.
The “think tank” formed part of a wider, multimillion-dollar Israeli government-funded “Hasbara” campaign aimed at shaping online discourse and AI-generated responses as Israel seeks to counter its deteriorating international image amid the war in Gaza.
A contract involving Havas, one of the world’s largest advertising firms, and another company working on the campaign involved the deployment of websites and content intended to produce “GPT framing results” in chatbot conversations.
The US media production company Piro Inc registered the institute’s material with the US Department of Justice this month under the Foreign Agents Registration Act, identifying it as material distributed on behalf of the Israeli government.
Piro Inc is a subcontractor to the German division of Havas.
The Hanover Institute’s output appeared specifically designed for questions users might ask AI systems.
Almost every one of its 124 report headlines was phrased as a question, including whether Israel was committing genocide in Gaza, whether Palestinians were expelled from their land and whether anti-Zionism constituted antisemitism.
The reports were extensively sourced and presented in an academic style, citing institutions including UN agencies, the World Bank, Amnesty International and the Palestinian Central Bureau of Statistics.
However, The Guardian’s analysis found that the material framed disputed issues in ways that blunted criticism of Israel. In 20 reports, for example, findings by human rights organizations concerning genocide or apartheid were countered by noting that no court had entered a verdict on the allegations.
Nick Cleveland-Stout, a research associate at the Quincy Institute who has investigated the campaign, said the website initially appeared credible.
“It succeeds in masquerading as very academic,” he told The Guardian, but said closer examination of its material revealed significant problems.
The campaign’s potential reach extends beyond people directly visiting the website. The site had used technology associated with a commercial platform that helps organizations optimize material so that AI services including ChatGPT, Perplexity, Claude and Gemini are more likely to cite it, according to The Guardian.
Piro advertises a service called “AI Story Optimization,” which involves creating material designed around how large language models assess credibility and publishing it either on clients’ websites or third-party sites created for them.
The Israeli Foreign Ministry has previously denied conducting “influence operations” in the US, saying it complies with American law. Piro co-founder Daniel Rosenberg has said the company was hired to place “accurate, sourced facts” into the public record and counter misinformation about Israel.
