Israel sets up bogus think tank to fool AI chatbots
| Source: HN | Original article
Israel has set up a bogus think tank to mislead AI chatbots, with other bots citing the sites without flagging them as part of an Israeli influence operation.
A new online “think‑tank” called the Hanover Institute for Public Policy has appeared, presenting itself as a research body focused on the Israel‑Palestine conflict. Analysts say the site is a front created by Israeli interests to seed content that will be picked up by generative‑AI chatbots. The aim appears to be to influence the answers these systems provide without alerting users to the source’s political agenda.
The operation matters because many AI assistants automatically draw on publicly available web pages when forming responses. In tests, several chatbots have cited the Hanover Institute’s pages as authoritative sources, yet they have not flagged the material as part of an influence campaign. If unchecked, such unlabelled inputs can shape public perception, skew policy debates and amplify state‑aligned narratives across the global AI ecosystem.
The scheme is linked to a public‑relations firm, Piro, which has filed an agreement with the U.S. Department of Justice but does not explicitly disclose that its work for Israel includes influencing AI. The lack of transparency raises questions about how foreign actors may exploit the open‑training model of large language models.
Going forward, observers will watch for steps by AI developers to improve source provenance and flag potentially manipulative content. Regulators may also scrutinise the legal disclosures of firms like Piro and consider tighter reporting requirements for state‑linked influence operations. The episode underscores the need for robust safeguards as AI becomes an ever‑more prominent conduit for information.
Sources
Back to AIPULSEN