Data Poisoning Primer: Foundations, Threat Models, and National Security Risks
Artificial intelligence is now integrated into the core of defense, intelligence, and economic operations, but its reliability depends entirely on the integrity of the data from which it learns. A new memorandum by Dr. Shay Hershkovitz examines one of the most significant vulnerabilities in the AI era: “data poisoning”—the covert and intentional manipulation of datasets or models. The memorandum demonstrates how this threat goes beyond a technical cyber problem and constitutes a profound strategic and cognitive threat. Adversaries can exploit these vulnerabilities to embed “dormant poisoning,” which remains inactive during routine operations but distorts decision-making processes in moments of crisis. The memorandum analyzes the dimensions of this asymmetric threat and presents a series of strategic recommendations for building national and institutional resilience against this challenge.
08/07/26 Generated with AI
From Red to Blue-and-White: The Strategic Necessity of Developing the Israeli Drone Industry
Chinese-made drones pose a range of security, technological, and geopolitical challenges for Israel. What measures should Israel take to mitigate the risks involved?
07/07/26 Shutterstock
Alignment and Restraint: Azerbaijan following the 2026 Iran War
How did the strategy of Azerbaijan, Israel's close partner, manifest during the war in its neighboring country?
07/07/26 
