IBM Study Finds Enterprises Locked Into AI Systems They Cannot Control

What Happened

IBM released findings from a new global study on June 17, 2026, showing that enterprises embedding artificial intelligence into core business operations face significant control limitations and rising vendor dependencies that leave them exposed to operational and strategic risk. The study, conducted by IBM's Institute for Business Value, found that most surveyed organizations are locked into AI systems they cannot easily modify, replace, or exit.

What the Study Found

According to the IBM Institute for Business Value report, the majority of enterprises surveyed reported limited ability to control the AI systems now running inside critical business functions. The study identified two primary risk factors: constrained technical control over AI system behavior, and growing reliance on a small number of external AI vendors and platforms.

The findings indicate that as AI becomes more deeply integrated into operations such as supply chain management, financial processing, and customer service, organizations lose flexibility to adjust or switch systems without significant disruption. The study did not specify the total number of organizations surveyed or the full geographic breakdown, but described the research as global in scope.

Background

IBM has been a long-standing provider of enterprise technology infrastructure and has increasingly positioned itself as a vendor of AI governance and hybrid cloud tools. The company's watsonx platform, launched in 2023, was designed in part to address enterprise concerns about AI transparency and control. IBM has published annual Institute for Business Value studies for more than two decades, using them to document shifts in enterprise technology adoption and risk.

The June 2026 report arrives as enterprise AI deployment has accelerated significantly across industries. Multiple major technology vendors, including Microsoft, Google, and Salesforce, have embedded AI capabilities directly into widely used enterprise software platforms over the past two years, reducing the friction of adoption but also tying organizations more closely to specific ecosystems.

What It Means in Practice

The study describes a structural exposure that emerges when AI systems become operationally critical before governance frameworks are in place to manage them. Organizations that cannot modify AI decision logic, audit outputs, or migrate to alternative systems face compounded risk if a vendor changes pricing, discontinues a product, or if the system produces errors in a high-stakes context.

IBM's findings align with concerns raised separately by regulators in the European Union, where the EU AI Act, which took effect in stages beginning in 2024, requires organizations deploying high-risk AI systems to maintain human oversight mechanisms and documentation of system behavior. Compliance timelines under that regulation continue through 2026 and into 2027.

The report does not name specific AI vendors whose systems are implicated in the dependency findings, nor does it rank industries by severity of exposure.

What Companies Said

IBM did not provide individual executive quotes in the initial wire summary of the study. The company framed the findings as a call for enterprises to prioritize what it described as AI systems offering greater transparency and portability, areas in which IBM has commercial products. The full study is available through IBM's newsroom.

What Comes Next

IBM has indicated it will present additional findings from the study at enterprise technology events scheduled for the second half of 2026, with a focus on recommended governance frameworks for organizations seeking to reduce AI system dependencies.