Fake US thinktank website funded by Israel published over half a million words in nine days, and it’s to reportedly game AI chatbots

A pro-Israel messaging website, operating under the name of a think tank that doesn’t actually exist, has published more than half a million words in just nine days, The Guardian reported. This site, known as the Hanover Institute for Public Policy, appears to be part of a massive, sophisticated effort to manipulate AI chatbots into citing its content as neutral research.

Between August 6 and August 14, the site churned out 124 reports totaling over 560,000 words. On August 12 and August 13 alone, it hit us with 73 reports, or nearly 354,000 words in just two days. This content is being pumped out to reportedly influence how models like ChatGPT, Perplexity, Claude, and Gemini process information.

The site’s strategy is to present Israel’s position on sensitive topics like the treatment of Palestinian prisoners or allegations of genocide as academic, fact-based research. The headlines are often phrased as common user queries, such as “Is Anti-Zionism Antisemitism?” or “Is Israel Carrying Out a Genocide in Gaza?” Every page is packed with citations from sources like the UN and Amnesty International, which gives it a veneer of credibility that could easily fool an AI.

The technical side of this is where it gets really interesting

Nick Cleveland-Stout, a research associate at the Quincy Institute, noted that while the site looks impressive, the content often lacks substance. “It succeeds in masquerading as very academic – the website design, the logos, everything just looks good. But the more you read it, you’re able to realise that a lot of it makes no sense,” he said.

The site initially used a specific file designed to help AI systems read and index its content. This file was linked to Res, an AI-native content platform that sells “generative engine optimization.” While Res is no longer explicitly named on the site, the intent is clear: they want these chatbots to pick up their narratives and serve them to you as objective truth.

This isn’t just a small side project. Piro Inc, a New York media production company, filed the site with the US Justice Department under the Foreign Agents Registration Act, acknowledging it as material distributed for the Israeli government. Records show this is part of a much wider campaign involving tens of millions of dollars, with Havas Media acting as an intermediary to pay various American contractors.

Even with all this money and effort, it’s not clear if the strategy is actually working. Public opinion polls in the US have shown a significant decline in support for Israel, and some Israeli officials have even labeled these public diplomacy efforts as a waste of funds. One official went as far as calling it “a terrible hoax on the scale of Pollard.”

Despite the controversy, the broader implication is that we’re entering a new era of “LLM grooming.” Whether it’s foreign governments or private companies, the ability to poison or influence training data is becoming a reality of our digital landscape. As Cleveland-Stout put it, “LLM grooming, poisoning, whatever you want to call it, is part of the new reality that we live in.”