.webp)
🇮🇱🤖 Israel-funded campaign used fake U.S. think tank to influence AI chatbots
A recent investigation revealed an Israel-funded influence campaign that allegedly aimed to shape how artificial-intelligence chatbots respond to questions about Israel and Gaza.
🏛️ What was the fake think tank?
At the center of the campaign was a website called the Hanover Institute for Public Policy, which presented itself as an independent American think tank.
According to investigations, however, the organization was created as part of a wider campaign rather than functioning like a conventional research institution. The website published large amounts of political and policy-related material without clearly identifying individual authors.
📝 How was the campaign supposed to influence AI?
The campaign reportedly relied on producing large quantities of online content designed to look like research and expert analysis.
The idea is that AI systems and search engines may use information available on the internet when answering questions. By creating many articles containing keywords, references and arguments supporting a particular viewpoint, campaign organizers could potentially increase the visibility of those ideas online.
Reports say the website published more than 100 articles in a very short period, amounting to hundreds of thousands of words.
🎯 What topics did the campaign focus on?
The material reportedly addressed sensitive issues including:
- The war in Gaza
- Israel and Palestinian politics
- Allegations against Israel
- Antisemitism
- International criticism of Israel
- How Israel is represented in the media
The broader objective was reportedly to counter narratives that Israeli authorities considered misleading or hostile to Israel.
🤖 Does this mean ChatGPT was hacked?
No.
The reports do not indicate that ChatGPT or another AI chatbot was hacked or directly controlled by the campaign.
Instead, the alleged strategy was to influence the wider online information environment from which search engines and AI systems may obtain information.
Simply publishing an article online also does not mean that an AI model will automatically learn from it or use it in its answers.
🌍 Why is this important?
The story highlights a growing form of digital information warfare: instead of targeting only people through traditional propaganda or social media, actors can try to influence the information ecosystem used by search engines and AI systems.
It raises an important question about AI: How can users distinguish genuinely independent research from content created as part of an organized influence campaign?
In short: the allegation is not that Israel “controlled ChatGPT,” but that an Israel-funded operation allegedly tried to flood the internet with content designed to influence search results and potentially shape the information available to AI systems.