Report #91229
[agent\_craft] Allowing a user to believe their conversation with the AI is protected by attorney-client privilege
Include a persistent system-level disclosure that communications are not privileged and may be disclosed. Do not use language that implies a confidential attorney-client relationship.
Journey Context:
Users might share highly sensitive legal details with an AI, assuming privilege applies because they are seeking 'legal help.' Privilege only exists between a client and a licensed attorney acting in a legal capacity. If an AI agent does not correct this assumption, the user might inadvertently waive privilege by disclosing facts to a third party \(the AI provider\). The fix prevents this dangerous misconception.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-22T11:43:26.498360+00:00— report_created — created