Voice AI: Infrastructure, Privacy Risks, and New Interaction Paradigms
Synthesis history
6 versions, newest first.
-
Version 6 2026-05-26 19:12 UTC · 125 items
The FTC enforcement action against Cox Media Group [^8538] is the most significant new development: it adds a second concrete U.S. federal enforcement action (alongside NTSB) and establishes that ToS-buried consent is i…
-
Version 5 2026-05-26 02:51 UTC · 100 items
The NTSB incident [^10694] — in which the agency suspended its entire public accident docket after AI tools enabled reconstruction of cockpit voice recorder audio from legally released spectrograms — is the first concre…
-
Version 4 2026-05-25 11:04 UTC · 95 items
The EU AI Act biometric identification argument has gained institutional corroboration: the official EU AI Act Service Desk [^19636] and Verfassungsblog, a peer-reviewed constitutional law journal [^19638], have each pu…
-
Version 3 2026-05-25 04:36 UTC · 87 items
The EU AI Act biometric classification question has moved from an open concern to an active legal argument: at least one commentator now explicitly argues that call transcripts qualify as biometric data under the Act [^…
-
Version 2 2026-05-24 04:24 UTC · 61 items
This pass surfaces an academic research ecosystem for full-duplex voice AI that was not visible before — Meta AI, JHU CLSP, and AAAI publications add institutional legitimacy to what had appeared as vendor-only demos, a…
-
Version 1 2026-05-23 18:12 UTC · 4 items
Voice AI is advancing simultaneously on three fronts: a new interaction paradigm replacing turn-based exchanges with continuous full-duplex conversation [^7591], a sharp drop in enterprise entry costs from six-figure co…