Descriptions:
Scott Galloway’s Prof G Conversations podcast features Meredith Whittaker, President of Signal, in a substantive discussion covering encryption mechanics, AI’s relationship with surveillance capitalism, and the structural privacy risks introduced by AI agents operating on personal devices.
Whittaker opens with a precise technical breakdown of what actually differentiates Signal from WhatsApp: while WhatsApp licenses Signal’s encryption protocol for message content, it does not encrypt metadata — who you contact, when, your contact list, group memberships, and profile photo. Signal encrypts the full stack, and as an open-source platform, allows independent code verification rather than requiring trust in the organization’s claims. She uses Signal’s public subpoena log at signal.org/bigbro to illustrate the gap between what governments expect to receive from a messaging platform and what Signal can actually produce.
The conversation expands into AI’s structural relationship with privacy, with Whittaker arguing that the dominant tech business model — collecting data via platforms to train models and sell targeted access — is incompatible with genuine privacy preservation. She applies this critique directly to AI agents: an agent planning a work dinner needs calendar access, browser access, and payment credentials, creating a significant data consolidation risk. Her concern about AI in high-stakes domains (nuclear, energy) is technically grounded rather than apocalyptic, distinguishing between reward hacking and emergent behavior risks versus more speculative existential framings.
📺 Source: The Prof G Pod – Scott Galloway · Published March 05, 2026
🏷️ Format: Interview
