“There Is No Legal Confidentiality When Using Chat GPT” Facing The AI Privacy Issue

Summary

The besties react to a clip of Sam Altman warning that conversations with ChatGPT carry no legal confidentiality and could be produced in a lawsuit. David Sacks explains that the law currently treats chat history like search history, which can be subpoenaed, but argues AI chats are far more personal and interactive, widening the scope for government intrusion. One bestie proposes letting AI models pass bar and medical certification so they would inherit professional privilege, which opens a broader question about rights and responsibilities for AI. Chamath argues that an encrypted-by-default 'Signal for LLMs' is a market opportunity and points to Grok, Apple's no-data stance and his own use of Brave.

• Sam Altman clip: there is no legal privilege for ChatGPT conversations, which could be produced in litigation.
• David Sacks: current law treats AI chat history the same as search history, which the government can subpoena.
• Sacks argues AI chats are more personal and interactive than search, so the intrusion risk is much greater.
• Proposal on the show: let AI models pass bar and medical certification and inherit professional privilege.
• Follow-on question raised about what privileged status would mean for AI rights and AI responsibility.
• Chamath: an encrypted-by-default AI chat product, the Signal equivalent of an LLM, is a market opportunity.
• Privacy defaults highlighted: Grok encrypting by default, Apple's 'I don't have it' posture, Brave browser and search.
• No public tickers, funds, commodities or securities are discussed, and no investment stance is taken in the clip.