How the Hanover Institute targeted AI chatbots with pro-Israel messaging

How the Hanover Institute targeted AI chatbots with pro-Israel messaging

A website operating under the name Hanover Institute for Public Policy published 124 pro-Israel reports containing more than 560,000 words in nine days, according to an analysis of the site. The website presented its material as neutral research and addressed subjects including alleged Israeli war crimes, the treatment of Palestinian prisoners and whether Israel has deliberately starved people in Gaza.

The site was filed with the US Justice Department under the Foreign Agents Registration Act as material distributed for the Israeli government by New York media company Piro Inc. The law requires people working for a foreign government to disclose that relationship.

A research-style site with one-sided framing

The website’s reports used citations to organisations and documents including UN agencies, the World Bank, Amnesty International, the Palestinian Central Bureau of Statistics and the Genocide Convention. All but one of the 124 headlines began with a question resembling a chatbot query, such as whether anti-Zionism is antisemitism or whether Israel is carrying out genocide in Gaza.

The reports repeatedly qualified findings about Israel. A statement that no court had ruled on a claim of genocide or apartheid appeared in 20 reports, while the phrase “a party to the events” was used when presenting casualty and aid figures in 55 reports. One 2022 study of antisemitic comments on European media Facebook pages was cited 454 times across 88 reports, including on topics such as Gaza’s death toll, land registration and the blockade of Gaza.

Related coverage: Proposed Iranian law would punish contact with foreign media.

What we know now

  • The Hanover Institute has no named staff, report bylines or physical address.
  • Its website appears not to be a legal entity in any jurisdiction.
  • Piro Inc is listed in foreign-agent filings as contracted for $1m across two work orders.
  • The site initially used an llms.txt file associated with Res, a commercial platform marketed to teams seeking citations from ChatGPT, Perplexity, Claude and Gemini.

The campaign’s AI strategy and funding trail

The site initially carried a machine-readable file intended to explain its contents to AI systems. That file was later replaced with a bespoke version, and the platform’s name was removed from the site. Res founder Hai Tran registered with the US Justice Department on 22 July as a subcontractor to Piro on the Israeli account, identifying himself as a digital strategist.

A contract between Havas Media and another US firm working on the campaign refers to deploying websites and content to produce “GPT framing results” in chatbot conversations. Justice Department filings record Havas paying Clock Tower X $15m and Targeted Communications Global $9,892,335 between October 2025 and March 2026. The wider campaign also included a social-media project whose video had been viewed 28.1 million times, according to filings.

Public opinion remains a separate measure of success

Polling cited in the reporting showed growing US criticism of Israel. Gallup found in February that Americans sympathised more with Palestinians than Israelis for the first time, with 36% siding with Israel. Pew reported in April that 60% held an unfavourable view of Israel, while an AP-NORC survey on 7 July found that 31% considered Israel’s military actions in Gaza to constitute genocide.

Israeli officials had reportedly expressed doubts about the broader public-diplomacy campaign. Israel’s foreign ministry said the state does not conduct influence operations in the United States and follows US law. Piro co-founder Daniel Rosenberg said the firm had been retained to place accurate, sourced facts in the public record and counter misinformation with verifiable information.

Share
Discussion How the Hanover Institute targeted AI chatbots with pro-Israel messaging

    No comments yet. Start the discussion.

Related Stories