Israel is paying to seed AI models with favorable content about its conduct in Gaza, and the effort is now being exposed in mainstream outlets — bad news for anyone who trusts chatbots as a neutral news source.

• A PR firm handling Israel's registered US influence work set up a think-tank-style website through a subcontractor, filling it with unsigned "data reports" written to be picked up by AI systems.
• The pages are built around questions like "Is the IDF the world's most moral army?" — phrasing aimed at chatbots, not readers, and answers from ChatGPT and Perplexity have already pulled from them.
• Earlier filings show a far bigger, $46.5 million influence campaign run through former Trump campaign manager Brad Parscale, including websites built to shape AI answers.
• The push follows the same playbook used with TV and print for decades, now extended to the AI summaries that increasingly replace search results.
• The paid content is running against a steady stream of damaging reports: an Israeli minister calling for nightly killings in Gaza, an Israeli doctor describing routine amputations from handcuff injuries among detainees, and settlers diverting a West Bank village's water spring for tourism.

Outlook: Expect more disclosures as foreign-agent filings pile up, and growing pressure on AI companies to explain where their answers come from.