AI Adoption Is Expanding. Is Enterprise Accountability Keeping Pace?

What Is Changing

McKinsey's 2025 global AI survey found that 88% of respondents reported regular AI use in at least one business function, compared with 78% the prior year. Most had not scaled AI across the enterprise; increased reported use should not be confused with full-scale deployment.

In a separate 2026 study conducted with EY, HFS Research surveyed 302 executives at Global 2000 enterprises. Only 17% reported clear, consistently applied accountability when AI-influenced decisions produced negative outcomes; 38% said accountability was handled case by case. On day-to-day governance, 32% relied on periodic reviews and approvals, while 10% reported continuous monitoring with automated escalation.

The studies do not measure a common rate of change. Their combined implication is a practical governance test: can management show current authority for an expanding AI use?

Why It Matters

Consider an illustrative inventory-planning assistant. The original approval permits transfer recommendations for one region. Later, the tool serves three regions and can initiate transfers below a newly configured threshold. The project's name and technology owner have not changed, but its operating permissions have.

A successful technical release does not establish whether the new authority was properly granted. Funding approval alone does not show who is now answerable for inventory availability, transfer costs or service disruptions. The expansion might fall within existing delegated authority. Management should be able to demonstrate that rather than assume it.

The governance gap arises when current permissions exceed the recorded decision boundary and the designated authority has not addressed the change.

The 6xD Reading

Within D2, Digital Cognitive Organization, decision authority must follow the business work that AI influences, even when a central technology team manages the platform. D4, Digital Transformation 2.0, connects authorization to investment sponsorship and measurable outcomes.

The 6xD Enterprise AI Governance System therefore calls for one current accountability evidence record per material use, connecting who decides, who pays and who owns the outcome. This is the 6xD application of the research, not a requirement claimed by the surveys.

Leadership Takeaway

At the next AI portfolio review, select one material use whose users, integrations or action rights expanded in the last quarter. Compare the latest approved scope with its actual permissions. Ask who authorized the expanded use, whether the change remained within delegated limits, who can intervene and which business executive owns the resulting outcome.

Where the record and operating permissions differ, assign a corrective owner and refer the expansion to the appropriate decision authority before further material changes, in line with the enterprise's approved thresholds.

Use the four-question test to verify AI decision authority →