AI Threat: Identity Security in Financial Services

An AI model recently cracked a 29-year-old bug in Squid proxy software. This technology allowed Unit 42 to reduce two years of pen testing into just three weeks. What's concerning is that organizations are estimated to be only three to five months away from these frontier AI models being weaponized. The bigger risk for banks isn't code itself, but identity security. Machine identities now outnumber humans 109 to 1, and this gap is growing. In fact, identity weakness was behind nearly 90% of Unit 42's incident-response cases last year. As Paul Leonhirth of Palo Alto Networks states, "zero trust is the lack of implicit trust," yet many banks still operate with too much implicit trust. This shift means AI has dramatically shortened vulnerability discovery timelines. Also, AI-enabled fraud is on pace to triple, potentially reaching 40 billion US dollars by 2027. This trend highlights the urgent need for robust identity and AI governance in financial services.