Source: Cloud Security Alliance
Author: unknown
URL: https://cloudsecurityalliance.org/articles/dlp-and-ueba-are-asking-the-wrong-insider-risk-questions
https://cloudsecurityalliance.org/articles/dlp-and-ueba-are-asking-the-wrong-insider-risk-questions
ONE SENTENCE SUMMARY:
Insider risk persists because DLP and UEBA detect discrete events, not contextual action sequences requiring investigations into intent and change.
MAIN POINTS:
- Malicious-insider breaches average $4.99M and take 260 days to identify and contain.
- Leaders agree insider risk matters, yet struggle describing effective, proactive insider-risk programs.
- Authorized, policy-compliant actions become risky only when viewed as a sequence over time.
- DLP is effective for compliance exits, but wasn’t designed to model insider intent.
- Rules can’t capture intent, sanctioned-tool abuse, temporal arcs, or non-data-moving misuse.
- DLP depends on correct labeling; retyped, pasted, screenshot, or new data often escapes controls.
- Shadow AI usage is surging; DLP logs can’t distinguish benign debugging from codebase exfiltration.
- UEBA detects deviations, but legitimate work frequently looks anomalous without business context.
- Alert queues drive tuning toward silence; rare-but-dangerous cases get suppressed as outliers.
- Both tools assume risks are specifiable מראש; insider risk instead requires situational reasoning.
TAKEAWAYS:
- Prioritize sequence-based behavioral timelines over isolated policy violations or anomaly scores.
- Treat insider risk as an investigation workflow, not primarily a detection-and-alerting problem.
- Add organizational context (role changes, approvals, resignations) to interpret identical telemetry differently.
- Address sanctioned-channel misuse and AI-driven data handling as first-class insider-risk vectors.
- Measure success by completed, actionable investigations shared across Security/HR/Legal, not alert volume.