A new cybersecurity report released by DNSFilter highlights that unapproved and unmonitored “shadow AI” usage has already triggered actual security incidents in roughly 20% of surveyed organizations. Driven by rising enterprise risks, 74.5% of firms reported larger cybersecurity budgets compared to the prior year, with AI security and governance emerging as the top spending priority for 40.3% of respondents. The study uncovered a massive governance gap: nearly 43% of organizations allow employees to grant third-party AI tools OAuth or API access to corporate systems before notifying IT or without any notification at all. While over half of firms claim to use AI governance or monitoring controls, telemetry from DNSFilter network data shows that roughly one in seven organizations connects to more than 10 distinct AI tools on a normal business day vastly exceeding executive awareness. The findings emphasize that increased spending on static policies fails to secure environments where non-technical employees independently authorize AI integrations, leaving organizations vulnerable to data leakage and unauthorized access.

Full research : New study shows one in five organizations has suffered a security breach from unapproved AI connections as decentralized OAuth authorizations outpace IT visibility.