Wiz warns AI adoption magnifies misconfigurations

ITPro reports that new research from Wiz contends that artificial intelligence is not conjuring novel categories of vulnerabilities but rather aggressively multiplying the real estate where classic, preventable errors, like credential leaks and misconfigurations, can wreak havoc.

The analysis revealed that a staggering eighty percent of cloud breaches last year stemmed from basic hygiene failures rather than sophisticated zero-day exploits, with threat actors dedicating more than half of their pre-attack efforts to simple reconnaissance and environment mapping. "What changed was not the existence of these risks, but the environments in which they appeared and the speed at which they could be exploited," the firm noted. As over eighty-five percent of organizations adopt AI services tightly coupled to sensitive data stores, the blast radius of a single exposed secret expands dramatically.

Meanwhile, adversaries are weaponizing AI not to replace tried-and-true intrusion methods but to accelerate and automate them, using generative tools to streamline internal discovery once a foothold is established. Wiz urges defenders to reframe that pre-compromise reconnaissance phase as a critical detection opportunity, emphasizing that continuous visibility into exposure and trust boundaries offers a narrowing window to disrupt attacks before they escalate.