An Israel-funded think tank, the Hanover Institute for Public Policy, has been accused of mass-producing "research reports" in an effort to get its material surfaced and cited by AI tools including ChatGPT and Perplexity.
According to BlockBeats, citing Beating AI, Hanover has published more than 100 articles on Israel-Palestine issues since Aug. 6. The pieces carry no author bylines, and many of the headlines are written as direct questions that users might type into an AI chatbot.
Content packaged for machine reading
The report said Hanover also provides an llms.txt file, organizing the site’s content into an index that is easier for machines to read. That structure can make pages simpler for large language models to process.
Politico test found repeated citations
Politico tested the systems with neutral questions and found that both ChatGPT and Perplexity repeatedly cited Hanover.
Media outlets sampled the articles
404 Media and Responsible Statecraft separately used AI detection tools to spot-check articles on the site. Most of the sampled pieces were judged to be AI-generated, according to the report.
PR firm openly markets optimization for LLM citations
The report also identified Piro as the public relations firm behind the effort. Piro publicly offers a service called "AI Story Optimization," which analyzes Reddit, YouTube, forums, media outlets, and other pages to determine which sources are most likely to influence model answers, then creates content designed to be cited more easily by LLMs.
Piro’s website is now directly using Politico’s test results to advertise that service.
FARA filings show project and payments
U.S. Foreign Agents Registration Act filings show that Piro worked through Havas Media on a communications project for Israel’s Government Advertising Agency, LaPam. Two payments disclosed in public filings total $1 million.