Gartner Warns CIOs Against Unchecked AI Adoption as Costs and Risks Rise

Organisations are investing heavily in AI, but turning that investment into measurable business results remains a challenge. Gartner has urged CIOs to take a more measured approach to AI adoption, balancing business value with rising costs, security concerns and governance gaps.

At the opening keynote of the Gartner IT Symposium/Xpo 2026, analysts highlighted how uncontrolled AI usage could undermine the very benefits businesses are looking to achieve.

Daryl Plummer, Distinguished Vice President Analyst at Gartner, stressed the importance of introducing stronger controls around AI adoption. These could include stricter data usage policies and limiting access to certain AI models.

Gartner’s research shows that only 20% of AI projects succeed, while 58% of executives report positive returns on their AI investments. However, achieving those returns is becoming increasingly expensive. According to Plummer, organisations are spending 18 times more to achieve successful AI outcomes than they did two years ago.

The growing demand for AI is also pushing up overall technology spending, despite falling token costs. This makes selecting the right projects and tracking actual business returns increasingly important.

AI Governance Struggles to Keep Pace

The gap between AI adoption and governance is another concern. Gartner estimates that 71% of CIOs plan to deploy agentic AI within the next 12 months, but only 21% of organisations currently have AI governance frameworks in place.

Kristin Moyer, Distinguished Vice President Analyst at Gartner, introduced the Value-Risk Flywheel, a framework designed to help organisations manage the relationship between business value and AI-related risks.

The framework highlights the need for centralised decision-making involving technology, security, investments and safety. Gartner recommends establishing a dedicated governance structure with representation from across the organisation to ensure that AI initiatives remain aligned with business priorities.

Unnecessary AI Usage Is Adding to Costs

Gartner also raised concerns about businesses using AI where it adds little value. While technology vendors have reported productivity improvements of 50% or more, actual productivity gains recorded by Gartner stand at 16%.

Poor-quality AI-generated output is adding to the problem. Around 40% of workers reportedly received AI-generated work slop within a month, with each instance requiring approximately two hours of additional work to correct. For an average S&P 500 company, this could result in annual losses of $52 million.

Plummer urged organisations to move away from measuring AI adoption through token consumption and focus instead on the value delivered. CIOs were encouraged to prioritise meaningful use cases and introduce clear guidelines to prevent unnecessary spending.

Security Risks Extend Beyond External Threats

Shadow AI is creating another challenge for enterprises, particularly when employees introduce unauthorised tools or experiment with AI features within existing software.

Gartner analysts pointed out that even well-intentioned employees can unintentionally expose organisations to security risks. This makes governance as important as conventional cybersecurity measures.

As agentic AI becomes more widely deployed, CIOs will need greater visibility into how these systems are accessed and used.

Gartner’s broader message is clear: successful AI adoption requires discipline. Organisations must focus on projects that deliver tangible results, control unnecessary consumption and establish governance before expanding deployment.

0 replies on “Gartner Warns CIOs Against Unchecked AI Adoption as Costs and Risks Rise”

Related Post