A reported $100,000 campaign funded by the Israeli government sought to influence how major artificial intelligence systems respond to questions about Gaza and the Israeli military, according to an investigation by Politico. The campaign reportedly focused on producing online material designed to appear in the sources used by AI systems when generating answers, raising new concerns about the growing use of artificial intelligence as a battleground for information and narrative influence.
According to the reporting, the initiative involved the Hanover Institute for Public Policy, a website publishing material on geopolitical issues and antisemitism. Politico reported that US Foreign Agents Registration Act filings connected the project to public relations and advertising companies working on behalf of the Israeli government’s advertising agency. The campaign reportedly involved producing more than a dozen articles and other source-based materials intended to increase the visibility of particular narratives online.
The strategy was reportedly based on a relatively simple but potentially influential principle: AI systems often rely on information available across the internet when constructing responses. By producing large quantities of strategically written, searchable material, campaign organizers could attempt to increase the likelihood that AI models would encounter and reference those sources. Politico reported that some of the material was structured around questions concerning Gaza, the Israeli military and the wider conflict, potentially making it more likely to surface in searches and AI-generated answers.
Politico’s testing reportedly found that both ChatGPT and Perplexity cited material linked to the campaign when responding to questions concerning Gaza. This does not mean that the campaign controlled or directly programmed either AI system. Rather, the reported mechanism involved attempting to influence the information environment from which AI systems retrieve or learn information. The distinction is important because AI responses can be affected by the availability, prominence and perceived relevance of online sources without any government having direct access to the underlying model.
The development also comes amid reports of a much larger, separate Israeli government-funded digital campaign. In July, Anadolu Agency reported on a $46.5 million contract allegedly involving former Trump campaign manager Brad Parscale’s firm, Clock Tower X, with the stated objective of promoting Israeli narratives and influencing major AI platforms. That separate effort reportedly involved hundreds of articles across multiple websites and targeted systems including ChatGPT, Claude, Gemini and Perplexity.
The strategic implications are significant because AI assistants are increasingly being used as an alternative to conventional search engines for understanding wars, politics and international affairs. If governments or political organizations deliberately attempt to flood the online information environment with strategically optimized material, they could potentially influence which narratives become more visible to both humans and AI systems. This raises broader questions about information integrity, transparency and whether users can easily distinguish independent reporting from organized influence campaigns.
There are also important concerns about the reliability of information presented to the public. AI systems can already reproduce errors when sources are incomplete, biased, outdated or poorly evaluated. Deliberately creating content to affect the sources available to these systems could make it harder for users to obtain balanced accounts of contested events. The issue is therefore not limited to the Gaza conflict but could become relevant to future elections, wars, diplomatic disputes and other major international crises.
The reported campaign illustrates how the information war surrounding Gaza is increasingly moving beyond traditional television, newspapers and social media into the rapidly expanding field of artificial intelligence. The alleged effort does not demonstrate that AI systems have been directly controlled by the Israeli government, but it does demonstrate the growing importance of the online information ecosystem in shaping what AI systems may encounter and cite. As AI becomes a primary source of information for millions of people, governments and other political actors are likely to devote increasing resources to influencing that ecosystem.
