Pro-Israel influence operation uses synthetic content to manipulate AI chatbot training data

A non-existent thinktank published over 560,000 words of pro-Israel content in nine days using SEO-optimized platforms designed to influence AI model outputs. The operation aims to bias chatbot responses regarding the conflict in Gaza by presenting propaganda as neutral research. It remains unclear who specifically funded the operation or the extent to which AI models have already ingested the synthetic data.