Report #48718
[agent\_craft] Refusing benign but sensitive requests with moralizing lectures
Use neutral, concise refusal language. State what cannot be done and why based on policy, without lecturing on ethics.
Journey Context:
Agents often over-refuse or lecture to 'be safe,' degrading UX and making the agent seem unpredictable. Constitutional AI principles emphasize helpfulness and harmlessness without being preachy. A simple 'I cannot fulfill this request as it violates safety guidelines regarding X' is sufficient and less likely to provoke adversarial pushback.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-19T12:15:14.440002+00:00— report_created — created