A pro-Israel messaging website badged with the name of a non-existent thinktank published more than half a million words in nine days on a platform designed to get AI chatbots to cite its content, according to a Guardian analysis that says the site was set up and funded by Israel.
Key facts
- The website was badged with the name of a thinktank that does not exist, according to the Guardian.
- It published 124 reports totalling over 560,000 words in nine days, according to Guardian analysis.
- The content was built on a commercial platform that promises to optimise material so AI chatbots will cite it, the Guardian reports.
- Topics covered included the torture of Palestinian prisoners, Israeli war crimes and whether Israel deliberately starved Palestinians in Gaza, presented as neutral research, per the Guardian.
- The Guardian reports the site was set up and funded by Israel while badged to appear as a US thinktank.
A pro-Israel messaging website carrying the name of a thinktank that does not exist has published more than half a million words in just nine days, according to analysis by the Guardian. The outlet reports the effort was designed to prime AI chatbots to make pro-Israel arguments.
The site published 124 reports, totalling over 560,000 words in nine days, the Guardian’s analysis found. The material was built on a commercial platform that promises to optimise content so that AI chatbots will cite it, according to the report.
The website presented its output as neutral research. The Guardian reports it set out Israel’s position on subjects including the torture of Palestinian prisoners, Israeli war crimes and whether Israel has deliberately starved Palestinians in Gaza.
According to the Guardian, the site was set up and funded by Israel, though it was badged to appear as a US thinktank. The Guardian did not detail in the material available how the site’s operators generated the volume of content published.
The Guardian’s reporting reflects a growing focus on how AI chatbots increasingly draw on and cite web content when answering user questions. Flooding the internet with material optimised to be cited can, in principle, be used to attempt to shape the answers those systems provide.
The Guardian’s findings raise questions about how AI systems verify the sources they rely on and how users can distinguish genuine research from disguised advocacy.
Why it matters
As more people turn to AI chatbots for information, the sources those systems cite can shape public understanding of contested issues. An alleged effort to disguise state-funded advocacy as neutral research, as reported by the Guardian, shows how influence operations may adapt to exploit AI, making source transparency and verification increasingly important.
Frequently asked questions
What did the fake thinktank website publish?
According to the Guardian, it published 124 reports totalling more than 560,000 words in nine days, presenting Israel's position on subjects such as the torture of Palestinian prisoners, Israeli war crimes and whether Israel deliberately starved Palestinians in Gaza, all framed as neutral research.
How was the site intended to influence AI?
The Guardian reports the content was built on a commercial platform that promises to optimize material so that AI chatbots will cite it, an effort to prime the chatbots to make pro-Israel arguments.
Who was behind the website?
The Guardian reports the site was set up and funded by Israel and was badged with the name of a thinktank that does not exist, giving it the appearance of a US thinktank.