Israel is trying to manipulate the output of artificial intelligence on Gaza

On 15 August, Israel launched a public relations campaign aimed at guiding the narrative of artificial intelligence systems such as ChatGPT in their responses to questions relating to Israel, Gaza and the Israel Defense Forces. The Hanover Institute of Public Policy, established by a number of Israeli-commissioned public relations companies, has issued numerous inquisitorial reports, such as “Is the Israel Defense Forces the world's most moral army?” “Is there a deliberate policy of creating famine in Gaza?” These papers are part of an Israeli-funded public information project with a total investment of $100,000. Tests show that when users ask questions about neutrality, the ChatGPT and Perplixity artificial intelligence model refers to information published by the Institute. Israel has also funded another public opinion intervention project, totalling $46.5 million, which is dedicated to the development of various web sites, output customizations and thus to the “content output orientation” of the GPT model。