Gartner Addressed AI Creep Risks at Conference
The firm highlighted how rapidly evolving vendor AI tools are challenging traditional corporate risk assessment models.
Updated on Sept. 24, 2026 in Artificial Intelligence

Live Poll
Do you trust that companies inform you whenever they update software with new AI features?
At the 2026 Enterprise Risk, Audit & Compliance Conference, Gartner identified the growing challenge of AI creep. This occurs when third-party vendors introduce new artificial intelligence capabilities into products without informing their customers.
Why it matters
Traditional assurance models, which rely on slow, periodic vendor disclosures and attestations, have become unreliable due to the rapid pace of change. Because agentic AI capabilities evolve faster than oversight protocols, companies are now forced to adopt continuous monitoring.
Traditional assurance models assume that vendor capabilities change slowly and rely on manual contractual disclosures. In contrast, new agentic AI features in third-party products update rapidly, rendering previous point-in-time assessment tools technically obsolete.
The players
Gartner
Gartner is a global research and advisory firm that provides information, advice, and tools for leaders in IT, finance, and other major business functions.
The details
IT and risk teams are increasingly shifting their focus toward maintaining continuous visibility into vendor product changes to mitigate unseen risks. This transition seeks to replace antiquated trust-based systems with dynamic evaluations that can track the deployment of autonomous AI agents in real time.
Timeline
The Gartner Enterprise Risk, Audit & Compliance Conference took place in 2026.
The Tech Race
This shift represents a fundamental pivot from manual, document-heavy auditing to real-time, automated verification processes. It pits agile risk-management strategies against legacy systems that were designed for a slower era of software updates.
Business leaders must now demand greater transparency from software vendors regarding hidden AI features that could introduce vulnerabilities. Organizations that fail to move toward continuous oversight may face increased operational risks as vendors deploy unvetted agentic capabilities.
The takeaway
Enterprises should treat all third-party software updates as potential security events rather than passive service improvements. Proactive, continuous monitoring is the only way to maintain control over an environment where AI features can be enabled at the flip of a switch.
Further reading
For more on the evolving standards for automated systems, visit our Artificial Intelligence section.
Source note: This article includes information reported by TechTarget.
Live Poll
Do you trust that companies inform you whenever they update software with new AI features?



