Israel Uses 'Generative Engine Optimization' to Shape AI Narratives on Gaza

The Israeli government has launched a strategic influence campaign to manipulate the responses generated by Large Language Models (LLMs) when users inquire about Gaza and the Israel Defense Forces (IDF). By utilizing a technique known as "generative engine optimization," the state sought to ensure that AI chatbots retrieve and cite favorable narratives as authoritative facts.

At the center of the operation was the Hanover Institute for Public Policy, a platform used to publish AI-optimized "data reports" on sensitive geopolitical issues. These reports addressed contentious topics, including whether the IDF is "the world's most moral army" and whether Israel is pursuing a policy of starvation in Gaza. According to FARA filings with the U.S. Department of Justice, a $100,000 campaign was established to disseminate "factual and source-supported material" to the American public through the institute. The initiative was funded by the Israel Government Advertising Agency (LaPam) and executed through Havas Media and its subcontractor, Piro Inc.

The Shift to Algorithmic Influence

This campaign marks a pivot from traditional Search Engine Optimization (SEO), which targets human clicks on a results page, to Generative Engine Optimization (GEO), which targets the training and retrieval data used by AI assistants. As global scrutiny of Israel's actions in Gaza intensifies and public support fluctuates in the United States, the Israeli government has expanded its public relations efforts to target the digital information ecosystem that millions of users now rely on for political and historical explanations.

Daniel Rosenberg, co-founder of Piro Inc., stated that the company was retained by Israel to "place accurate and sourced information in the public record and counter what it regarded as misinformation about the country."

Implications for Digital Truth

This strategy represents a new frontier in political warfare, where states attempt to manipulate perceived truth at the algorithmic level. Because AI systems synthesize multiple sources into a single, authoritative-sounding answer, state-funded content optimized for these models can bypass the transparency typically associated with traditional advertising. This creates a risk where users believe they are receiving neutral, synthesized information while actually encountering state-sponsored advocacy.

As Arab News noted in its analysis, the objective of such campaigns is no longer merely to persuade the person reading the answer, but to influence the information environment from which the machine constructs the answer in the first place.

Evidence of Impact

The effectiveness of the campaign has already been observed in real-world testing. Tests conducted by Politico revealed that both ChatGPT and Perplexity cited material from the Hanover Institute when answering user questions regarding Gaza, anti-Zionism, and antisemitism. These results suggest that the AI models viewed the institute's optimized content as a reliable source for their synthesized responses.

Observers now watch to see if other nations will adopt similar GEO tactics to shape global perceptions, as the boundary between organic information and state-sponsored algorithmic influence continues to blur.