Can enemies hack military AI? Iron Dome engineer on battlefield risks

Col. (Res.) Ryan Gity, who helped develop Iron Dome and shaped Israel's national AI strategy, reveals how adversaries can plant "zero-day triggers" in open-source AI models used by militaries worldwide—including the IDF and U.S. government. In this wide-ranging conversation, Gity explains why "AI is the next nuclear arms race" and shares a striking anecdote from a Geneva ethics conference where a Russian representative declared they would use autonomous weapons regardless of Western concerns. He discusses the recent discovery that 80% of the world's 1.5 million AI models are open-source and potentially compromised, why Israel needs its own sovereign foundation model to protect against historical manipulation, and how robotics and AI could solve the IDF's manpower shortage. Gity offers a rare insider perspective as someone who served on two committees appointed by the Prime Minister—one on defense budgets and future forces, another on Israel's national AI strategy. He addresses the ethical frameworks governing autonomous weapons, the chip supply chain battle between the U.S. and China, and predicts humanoid robots in homes within 15 years.