Last week a legal case tacked a novel issue involving Artificial Intelligence. The defendant used Claude, Anthropic’s AI platform, to prepare reports that outlined his possible defense strategy. The defendant argued that the reports were subject to privilege. The government argued they were not and wanted them produced for discovery in the legal case.
The government won.
Judge Rakoff in the Southern District made the ruling. The court’s reasoning
- First, the reports were not part of any attorney-client relationship. Claude is not an attorney.
- Second, the defendant did not communicate with Claude to obtain legal advice. When the government asked Claude whether it could give legal advice it returned: “I’m not a lawyer and can’t proved formal legal advice or recommendations.”
- Third, the communications memorialized in the AI reports are not confidential. Anthropic’s privacy policy states that it collects data to train Claude and reserves the right to disclose data to third parties, including governmental regulatory bodies.
More detail in in the Debevoise Data Blog and court ruling.
Sources:
- SDNY Rules AI-Generated Documents Are Not Protected by Privilege By: Charu A. Chandrasekhar, Avi Gesser and Caroline Wallace in the Debevoise Data Blog
- The ruling is below