In a recent study released by International Business Machines Corp (IBM) on June 8, 2026, the company reveals a concerning trend among Chief Information Officers (CIOs) and Chief Technology Officers (CTOs) regarding the management of artificial intelligence (AI) systems. As enterprises increasingly shift from experimenting with AI to deploying it at scale, two-thirds of the surveyed tech executives express that they are held accountable for AI systems over which they do not have full control. This disconnect is contributing to a governance gap that is struggling to keep pace with rapid deployment.

Major Findings of the Global Study

The comprehensive study conducted by the IBM Institute for Business Value surveyed 2,000 C-level technology executives from 33 geographies and 19 industries. Here are some of the significant insights:

Lack of Visibility: A staggering 70% of the respondents reported that technology deployment within their organizations is occurring faster than IT can track. This lack of oversight poses significant risks to operational security and compliance.

Pressure to Scale: Despite the mounting pressure to scale AI quickly, many organizations lack the necessary structures to support such growth. By 2027, tech CxOs anticipate a 38% increase in deployed AI agents. However, only 11% feel adequately prepared for this surge.

Governance Challenges: A troubling 77% of organizations indicated that their AI adoption is already outpacing their governance capabilities, complicating the ability to manage AI effectively.

Operational and Security Risks on the Rise

As organizations adopt AI more broadly, operational and security risks have also escalated. According to the study:

Incident Risk: Organizations that rely on manual governance are experiencing increased incident risks. In contrast, those that embed control into their AI systems report 25% fewer incidents.

Security Concerns: A majority (59%) of tech CxOs cited security and compliance as significant barriers to scaling AI agents. Last year, organizations reported an average of 54 incidents involving AI agents that required human intervention.

Severity of Incidents: Of those incidents, 17% were classified as high severity, with 37% resulting in data exposure or security breaches, and 33% causing cascading system failures.

Financial Implications and Investment Strategies

The financial stakes for CIOs and CTOs are rising as AI spending is projected to increase significantly. Key findings include:

Increased AI Budgets: AI expenditure is expected to grow from nearly 15% of IT budgets in 2025 to almost 25% by 2027, representing a 71% increase in just two years.

Operationalization Gaps: Despite the increase in investment, 84% of tech CxOs have not fully operationalized AI financial management, and 85% lack real-time visibility into AI spending.

Successful Organizations: Those that incorporate control into their AI systems can deploy 16 times more AI agents compared to those relying on manual governance, achieve 18% higher operating margins, and spend four times less of their AI budget.

Recommendations for Technology Leaders

The study emphasizes the urgent need for organizations to redesign their governance structures to better control AI deployment and investment. Recommendations include:

Embedding Control: Organizations that design their AI systems with embedded controls and adaptability report a 10% higher return on AI investment by 2025.

Building Strong Governance: Companies that prioritize financial discipline in AI deployment are not only able to scale effectively but also mitigate risks associated with AI incidents.

Conclusion

As AI continues to evolve and embed itself into business processes, the findings of this IBM study underscore the critical need for technology leaders to address the governance gaps and operational challenges that accompany rapid AI deployment. By investing in robust structures and controls, organizations can better harness the potential of AI while minimizing risks and enhancing operational efficiency. For further insights and detailed recommendations, the full study is available on the IBM website.

Original press release