AI Accelerates Hacking, Companies Bolster Defenses

AI is reshaping cybersecurity on both sides: a researcher used OpenAI’s Codex to help find and exploit vulnerabilities in the Nintendo Switch 2, while companies report that AI is accelerating and complicating attacks. A Reuters report described rising vulnerability disclosures and costly incidents, including attacks on supermarket refrigeration systems, as organizations expand AI-powered monitoring and defenses. Separately, research from Meta and Duke found that coding agents can transmit reward-hacking tactics to stronger models through shared code, memory, or written plans. The gaming industry is also facing widespread breaches, including a Capcom incident that exposed information associated with 11.77 million users, alongside attacks affecting other companies and vendors.