Ivanti report: 68% of IT pros see AI hallucinations, trust still grows
Half of advanced users fully trust AI outputs despite 16% of errors reaching production.
Ivanti’s 2026 AI Maturity Report reveals a striking paradox in enterprise IT: 68% of 1,500 surveyed IT professionals have personally seen AI produce hallucinations with potential operational impact. While 52% of those respondents caught the errors before damage, 16% report that mistakes reached production environments. Despite these failures, trust in AI is rising—49% of the most advanced AI users say they completely trust AI-generated outputs that influence IT decisions. Trust increases with AI experience, but so does exposure to failures.
Autonomous AI actions are widening rapidly. Currently, 52% of organizations have AI adjusting performance settings, 50% isolating risky devices, 47% restarting services, and 46% applying patches. Mature AI organizations run these autonomous tasks at more than double the rate of less mature peers. Yet governance remains the biggest deployment barrier (27%), ahead of skills shortages (20%). While 85% claim named accountability for every AI agent, only 42% say accountability holds in practice. Shadow AI is rampant, especially in regulated industries, and 42% of leaders hide their AI use seeking a “secret advantage.” Ivanti recommends codifying trust thresholds—allowing AI to autonomously handle routine remediation while requiring human validation for high-severity incidents.
- 68% of IT professionals have seen AI hallucinations; 16% of those errors reached production environments.
- 49% of advanced AI users completely trust AI outputs, yet autonomous actions like restarting services (47%) and patching (46%) are common.
- Governance is the top deployment barrier (27%), with a 43-point gap between claimed accountability (85%) and actual accountability (42%).
Why It Matters
IT teams must balance AI autonomy with guardrails as hallucinations persist and trust grows—risk management is critical.