Israel Hires Firms to Influence AI Chatbots via Fake Think Tank

The Israeli Government Advertising Agency contracted Piro, Inc. to establish the Hanover Institute for Public Policy, a fake think tank designed to influence the outputs of large language models such as ChatGPT, Gemini, and Claude. Since August 6, the institute has published more than 100 reports on the Israel-Palestine conflict, utilizing academic formatting and a neutral tone to mimic a credible American organization.

This strategy, described as AI Story Optimization or LLM poisoning, aims to ensure AI chatbots incorporate specific pro-Israel narratives into their responses. Piro, Inc. received $900,000 for the project, which was subcontracted through Havas Media. While the content frequently cites Israeli government sources and is largely AI-generated, Piro co-founder Daniel Rosenberg claims the objective is to counter misinformation with verifiable facts.

The operation is part of a larger effort to shape digital narratives. This includes a separate $46.5 million contract with Brad Parscale to develop pro-Israel websites specifically engineered to influence chatbot evaluations of credibility.