The Israeli government-backed Hanover Institute for Public Policy has been churning out "research reports" at an industrial pace, with one goal: getting cited by AI chatbots like ChatGPT and Perplexity. Since August 6, the site has published more than 100 articles on Israeli-Palestinian issues, none with a named author and many with headlines written to look like questions a user might ask an AI. The institute even maintains an llms.txt file, a simple standard for handing a site's entire contents to machine readers.
The strategy appears to be working. When Politico tested with neutral queries, both ChatGPT and Perplexity repeatedly cited Hanover. Independent outlets 404 Media and Responsible Statecraft then ran the articles through AI detectors and found most were flagged as machine-generated.
The operation is run by a PR firm called Piro, which openly sells an "AI Story Optimization" service. It analyzes which pages on Reddit, YouTube, forums, and news sites most influence model answers, then creates content engineered to be cited by large language models. Piro's website is now even using Politico's test results as a marketing pitch.
US Foreign Agents Registration Act (FARA) filings show that Piro, working through Havas Media, handled communications projects for Israel's Government Advertising Bureau, known as LaPam. Two disclosed payments total $1 million.
Source: 404 Media