PR Battle Breaks Into ChatGPT: Israeli-Funded 'Think Tank' Mass-Steals AI Citation

Dynamic Beating AI News: The Israel government-funded Hanover Institute for Public Policy has been exposed for mass-producing "research reports," with one of the goals being to influence AI search results such as ChatGPT and Perplexity. Since August 6, the website has published over 100 articles on Israeli topics, all without author attribution, with many headlines framed as questions that users would ask AI. Hanover has also provided llms.txt, organizing the entire site's content into an index easily readable by machines.

When Politico conducted a neutral question test, both ChatGPT and Perplexity frequently referenced Hanover. 404 Media and Responsible Statecraft used AI detectors to sample articles, with the majority being identified as AI-generated.

Behind-the-scenes PR firm Piro openly sells "AI Story Optimization" services, analyzing which pages on platforms like Reddit, YouTube, forums, and media outlets have the most impact on model responses, and then creating content specifically tailored to be easily referenced by large language models (LLMs). Piro's website now even directly uses Politico's test results to promote its services.

U.S. FARA documents reveal that Piro, through Havas Media, conducted a communication project for the Israeli government's advertising agency LaPam. Two entries in the public records show a total of $1 million.