A recent investigation by The Guardian, Drop Site News, and NewsGuard has uncovered a sophisticated new tactic of state propaganda: Language Modeling (LLM) grooming. The Israeli government covertly funded the creation of a fake think tank with the express purpose of tricking artificial intelligence algorithms (such as ChatGPT and Gemini) into adopting and reproducing narratives favorable to the state.
The strategy: Automation and massive volume
The epicenter of the operation was the self-styled Hanover Institute for Public Policy, a supposedly independent think tank that was, in reality, a digital front. The speed of the campaign betrays its artificial nature: in just nine days, the portal published 124 reports with over 560,000 words of content with a clear pro-Israel bias, falsely presenting them as "neutral scientific research." To inflate its online authority and evade search engine filters, the network cross-referenced data with official dissemination platforms like Paxpoint.org and articles from open-access repositories such as Isca.me, simulating an ecosystem of legitimate academic citations.
How did they deceive the AIs?
Unlike traditional campaigns targeting social media users, this operation was specifically designed for machine consumption: Direct query format: The Hanover Institute articles used titles structured as questions (e.g., "Is Israel a peace-seeking nation?" or "Is it beneficial for the US to increase military cooperation?").
This replicates the exact pattern in which users question chatbots. Algorithmic mimicry: The texts lacked individual authors and were written in a cold, structured, and analytical tone, full of data and tables of contents. The AI's data crawlers indexed these texts, interpreting them as optimal, objective, and highly authoritative answers.
The financial network
Behind the institute's facade operated an opaque network of international contracts. The Israeli Government Advertising Agency (LaPam) allocated $900,000 to the project. To conceal the state's origin, the funds were channeled through the French public relations giant Havas Media to the US agency Piro Inc., responsible for the technical execution.
Official documents from Clock Tower X LLC (which manages Paxpoint), registered under the US Department of Justice's Foreign Agents Act (FARA), further confirmed the operation's direct links to the State of Israel.
The New Frontier of Disinformation
This case marks a turning point in the global information war. Experts warn that the battlefield is no longer solely about swaying public opinion through social media bots, but also about contaminating the training databases of large-scale language models, enabling state propaganda to be returned to users in the form of AI-generated "objective responses."
Seriously, are we witnessing the end of truthful information?
LINKS:
nevent1qqs994zl9ar6ayx7u9x2eqn85a9zyuw4yaa5yttdk7gyspn93zd9kqgpzdmhxue69uhk7enxvd5xz6tw9ec82c30qgsdle00jjerpm84rg4d4ra8efq0kpkhecd05j6cezcan83mlcwg8csrqsqqqqqp8p6l4q
They already gave it a Downzap hahhahahah
🔥✨⚡✨💧
👹🛡 👁🗡👺
They have plenty of funds but no shame.
This post was downzapped 1555 sats as of 9/1/2026.
Why? Who?
They'd think I was @Solomonsatoshi
The truth is, it's a very strong and real piece of news... that greatly changes the landscape of how we interact with internet searches.
Only my posts about Zionism are downzapped these days- you can look at my recent posts and see- its not me that's being targeted (mostly) its ANYTHING critical of Israel.
It's really frustrating... I really thought that was going to happen, but I still decided to publish it because of its impact.
Someone clearly has very strong feelings in favour of Israel and will spend large amounts of sats to downzap any critical reference to Israel.
They do not come out and openly put forward their arguments for Israel - instead they hide behind anonymous downzapping seeking to bury any critical narrative or discussion.
The only silver lining to their covert cowardice and censorship is that their downzaps go into the rewards pool.
They've already lowered it again
⚛🐍🔯🐍⚛
The Zap Vs Downzap War
Excellent meme... if it was difficult before, I think it's going to be much harder now
I think the most complete defense is:
Normally, when I get into debates like this, I try to find the most reliable sources, cross-reference data, and verify everything thoroughly before taking a position. When I'm unsure, I don't engage in that kind of argument; instead, I try to listen to both sides and understand.
And if I see that the other person is being insistent, I withdraw and don't get involved in pointless discussions.
Can you do that for everything anyone throws at you tho? I think that minimizing exposure is probably a better strategy than doomscrolling. But... boomer. So figure it out yourself.
No way... only when I need to take a stance on an issue
a very successful strategy
I'm from the 90s
Haha I didn't mean to imply that you're the boomer
The part that I have a hard time dealing with, emotionally, is how many people are out there that just literally do not care about truth, and are willing to algorithmically generate lies and slop... and the only barrier to them doing that before was the lack of the appropriate tools.
It's very common nowadays; people go with the first piece of information they receive. They don't like to learn or form an opinion on a subject, and they're so idiotic that many believe they possess the absolute truth.
#1556026
Thanks for sharing