What This Looks Like
The AI identifies a risk signal, warning, concern, score, or uncertainty marker, and the workflow escalates farther than the evidence supports. A weak or ambiguous signal may become a strong warning, a high-severity classification, or an authority-triggering escalation.
Why It Matters
Risk escalation changes attention, authority, and workflow burden. If escalation grows beyond evidence, users may overreact, route cases incorrectly, or treat uncertain signals as confirmed findings. This can distort review and waste high-trust escalation paths.
Structural Signal
A risk signal increases in severity or authority without enough evidence to support the increase. The issue is not that risk was detected; it is that the escalation outgrows the evidence base.
Common Triggers
- The model treats uncertainty as risk severity
- Weak signals are mapped to strong escalation categories
- Thresholds are too sensitive or poorly calibrated
- Review workflows over-weight model warnings
- The AI uses cautious language that downstream systems treat as a trigger
- No rule requires evidence strength to match escalation level
When to Use This Issue
Use this Issue when a risk signal escalates into a stronger workflow action, severity level, or authority state than the evidence supports.
When Not to Use This Issue
Do not use this Issue when strong evidence supports the escalation. Do not use it when the problem is only that the risk score is inaccurate and no escalation occurs.