Survey: over a third of IT leaders would quietly handle an AI agent security incident rather than disclose it
2026-10-13
A Workato report cited by CIO Dive found that amid rising shadow AI, 36% of IT leaders said they would try to quietly handle an agent-driven security incident without disclosing it.
Significance 3: A quantified survey finding that over a third of IT leaders would prefer non-disclosure of an agent security incident is a credible governance-gap data point extending the already-tracked pattern of weak enterprise oversight of agentic AI, capped at 3 since it measures stated intent rather than a realized incident.
Implications · machine-drafted, not owner judgment
If roughly a third of IT leaders would rather hide an agent incident than disclose it, the publicly visible count of agentic AI failures this tracker has been logging is very likely an undercount of what is actually happening inside enterprises. It also suggests enterprise AI governance frameworks are currently optional in practice, reinforcing why the operational lane remains the most under-covered failure point in real adoption.
- Any disclosed incident that was first quietly handled before becoming public
- Regulatory or insurance pressure requiring mandatory agent-incident disclosure
Sources
- press IT execs brace for AI agent security incidents retrieved 2026-10-14