Risk register
AI-specific risks with likelihood, impact, and mitigation. Review these before scaling any pilot.
Sensitive company or customer data exposed to public AI tools
Mitigation: AI use policy + data classification + approved tool list
AI hallucinations reach customers unchecked
Mitigation: Human review step before any external use
AI use violates industry or customer regulations
Mitigation: Legal review of AI use cases in regulated areas
Locked into one AI platform with no exit path
Mitigation: Multi-vendor strategy; keep data portable
Staff resistance or fear slows adoption
Mitigation: Change management plan + visible leadership support
Shadow AI — staff using unapproved tools without oversight
Mitigation: Approved tool list + network monitoring + policy
AI output is biased or unfair to groups
Mitigation: Bias testing + diverse review panel for key use cases
AI tool outage disrupts operations
Mitigation: Documented fallback process per AI use case