EY-Parthenon · September 2026

EY AI Risk and Governance Survey 2026

EY's AI Risk and Governance Survey 2026 (September 2026; n=202, executive survey) found that 98% have formal AI governance policies, yet 47% say their organization has previously not applied its AI governance process for urgent deployments.

Read the original report ↗Cite2 min read · Summary updated
Almost half have bypassed their own AI governance for an urgent deploymentPercent of senior AI decision-makers at US public companies with $1B+ revenue, 2026 (n=202). 53% is the remainder who did not report skipping it.
47%
  • Have skipped the governance process for an urgent deployment47%
  • Did not report skipping it53%
Source: EY, AI Risk and Governance Survey, 2026.

Key findings

  1. 01
    98% have formal AI governance policies in place, but 47% say their organization has previously not applied its governance process for urgent deployments.The respondents are senior AI decision-makers at US public companies with $1B+ revenue.
  2. 02
    91% say their organization uses agentic AI, and 85% of those admit at least a handful of these systems execute actions without real-time human involvement.
  3. 03
    26% of respondents whose organization uses agentic AI say it cannot detect unauthorized AI agents.
  4. 04
    About a third (36%) have experienced an AI incident or failure with a materially negative impact, such as data loss, financial damage or operational disruption.

By the numbers

98%have formal AI governance policies
47%have skipped the governance process for urgent deployments
36%experienced an AI incident with materially negative impact

What it means for you Draft

For executives at $10–100M companies

Having a policy and following it are different things, and this survey suggests the gap shows up under time pressure. For a $10–100M company, the question is whether an urgent deployment still gets a quick, real review. The sample is much larger companies than yours.

For practitioners

Agentic systems acting without real-time human involvement are common in this sample. Inventory your agents and decide how unauthorized ones would be detected.

Limitations

Medium trust.Disclosed sample (n=202, ±7 points, dated fieldwork) of large US public companies; EY sells AI consulting and assurance services.

Sample of 202 has a stated margin of error of ±7 points and covers only US public companies with $1B+ revenue. EY commissioned it through its assurance practice, and EY sells AI governance and assurance services.

About the publisher: Sells AI strategy and implementation consulting.

Cite the original

EY-Parthenon. "EY AI Risk and Governance Survey 2026." September 15, 2026. https://www.ey.com/en_us/newsroom/2026/09/ey-survey-finds-that-autonomous-ai-implementation-outpaces-oversight-yielding-an-ai-governance-gap