The country has launched a campaign to shape how AI models answer questions about its war on Hamas and the IDF amid declining public support in the US
Israel has spent tens of millions of dollars to boost its image in the US by seeking to influence how AI platforms portray the country, Politico has reported.
US public attitudes toward Israel have deteriorated in recent years, first over its response to the October 7, 2023 Hamas attacks and more recently over the US-Israeli war on Iran and its cost to taxpayers. Six in ten US adults now view Israel somewhat or very unfavorably, according to a June Pew Research Center poll, up from 53% last year and 42% in 2022. Israeli Prime Minister Benjamin Netanyahu is even less popular, with 65% expressing little or no confidence in his policies.
In a report on Friday, Politico detailed a $100,000 Israeli campaign aimed at influencing how large-language models such as ChatGPT answer questions about Gaza and the Israel-Hamas war. The effort centers on the Hanover Institute for Public Policy, a website publishing anonymous reports on anti-Semitism and geopolitical issues. The project was set up through French PR firm Havas Media, which handles much of Israel’s US influence work registered under the US Foreign Agents Registration Act (FARA), and its subcontractor, ad agency Piro Inc, on behalf of the Israel Government Advertising Agency (LaPam).
The Hanover Institute describes its work as “documentary, not declarative” and says its research is purely academic. Politico, however, cited FARA documents filed with the US Justice Department last week linking more than a dozen of its articles to a PR campaign that involved “the creation and dissemination of factual, source-supported informational materials” intended to “educate” the US public about Israel.
The website hosts multiple unattributed reports framed as questions, including “Is Torture of Palestinian Prisoners a War Crime?” and “Who Is the Aggressor in the Israel-Palestine Conflict?”, which Politico said suggests the material was designed to be picked up and cited by LLMs. The outlet also said technical markers linked the website to Res, a Seattle-based AI startup that helps clients get recommended by AI models. The link was confirmed by Res CEO Hai Tran in an email.
While Piro’s FARA filings do not explicitly identify influencing LLMs as a goal, Politico said both ChatGPT and Perplexity cited Hanover Institute material in response to neutral queries about Gaza and Israel.
It is not Israel’s first Havas-linked effort to sway public opinion, including through AI. According to a WSJ investigation published last month, the firm previously hired former campaign manager for US President Donald Trump, Brad Parscale, for a $46.5 million campaign that included pro-Israel AI-generated messages targeting Americans, as well as websites and posts designed to generate more favorable responses from AI platforms. Last fall, Responsible Statecraft reported that Havas Media also managed an “Influencer Campaign” which paid $900,000 for producing Israel-friendly content.
US frustration with Israel has extended beyond the general public, with politicians also becoming increasingly critical. Last month, more than 100 House Democrats backed an amendment that would have cut $3.3 billion in US aid to Israel. The measure failed due to Republican opposition but highlighted a growing partisan divide: just over two years ago, only 37 House Democrats backed a similar effort.
Trump has been largely supportive of Israel but has repeatedly criticized Netanyahu, who slammed the ceasefire with Iran announced by the US president earlier this summer and rejected his 15-point Board of Peace plan for Gaza. Trump also has yet to endorse Netanyahu ahead of the October election and has repeatedly dodged questions about his re-election, according to media reports. Two senior US officials told Axios this week that while Trump likes Netanyahu personally, he has grown frustrated with his policies and called Netanyahu “his own worst enemy.”
