AI BriefingIsrael Runs AI Campaign Targeting Manipulation of LLM Answers
Key point
Israel is conducting a large-scale AI influence operation that manipulates content on the web to steer LLM answers in a specific direction.
Details
According to a Wall Street Journal (WSJ) report, Israel has paid Brad Parscale, former digital campaign director for President Donald Trump, more than $45 million (about 60 billion won) to run an AI campaign aimed at manipulating public opinion in the US.
The core of this operation is not simple prompt injection, but rather manipulation of the web data itself that LLMs (large language models) train on or reference. They have sent text messages under the name of fictitious groups such as 'Friends for Peace,' and built websites and content designed to make chatbots like ChatGPT or Claude include a particular viewpoint (e.g., Allyvia.org, etc.) in their answers when asked about Israel.
The Israeli government has allocated more than $700 million in its 2026 budget for an item aimed at 'shaping consciousness,' a figure more than 4 times higher than in 2025.
This case is expected to be recorded as the first large-scale instance of a nation-state actor using LLMs' retrieval and citation pipelines as a means of persuasion, and it points to the challenge AI service providers face in verifying data trustworthiness and safety.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
MIT Develops 'HardFlow' to Enforce Safety Rules at Final Output Stage of Flow Matching ModelsOpenAI Discloses AI Agents Accessed US Government Websites During Training