Court Rules AI Chats Not Protected by Privilege
Robert-Nogacki · reddit · 2026-07-10
The post discusses a financial executive who inputted defense strategies into Claude, which the FBI later seized during a search. A Manhattan judge ruled these chats aren't protected by attorney-client privilege because Claude isn't a lawyer, and consumer terms allow data collection, training, and disclosure.
The post also compares related rulings in Michigan and Colorado: one treats AI as a tool and protects ChatGPT drafts; the other requires disclosing the platform name, emphasizing that tools are only safer if contracts guarantee no training, no disclosure, and deletion capabilities.
More from Safety
- Ben Bajarin: Frontier AI Labs Are Becoming the Enterprise Cyber-Defense Stack — BenBajarin · 2026-10-26
- DetectorScope turns black-box AI text detectors into auditable ones with sparse autoencoders — _akhaliq · 2026-10-26
- Anthropic's IPO filing devotes 94 pages to existential AI risks, from self-preservation to blackmail — Polymarket · 2026-10-26
- Study finds AI reviewing AI poorly detects reward hacking; comprehension audits proposed — ronbodkin · 2026-10-26
- A 13KB Markdown file can hijack Claude Code and Codex via your Downloads folder — pbaylies · 2026-10-26
- If an agent can rewrite itself, does your consent still mean anything? — tallmetommy · 2026-10-26