Pindrop traces the Biden New Hampshire robocall to ElevenLabs

AI system

System: Pindrop deepfake audio analysis (voice-engine attribution)
Model: Not assessed
Release: Not assessed
Application developer: Pindrop
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: A clean audio sample of the deepfake. A library of fingerprints for many commercial voice-synthesis engines. A second, independent model run on the same sample.
Human oversight: Not assessed
Relation to AI: The attribution is the vendor's own analysis, reported by WIRED with named sources and independently replicated by Farid's team. ElevenLabs did not comment on the incident. The sources document no harm stopped and do not link the later charges against Kramer or the FCC fine to Pindrop's finding.
AI contribution: The attribution is the vendor's own analysis, reported by WIRED with named sources and independently replicated by Farid's team. ElevenLabs did not comment on the incident. The sources document no harm stopped and do not link the later charges against Kramer or the FCC fine to Pindrop's finding.

Participants: Developer: Pindrop. Operator: Pindrop (developer and deployer are the same company); independent corroboration by Hany Farid's UC Berkeley model.
Jurisdictions: Developer: United States. Operator: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: The generating engine must be among those fingerprinted. Sample length and quality sufficient for attribution.
Replication attempts: Not assessed
Later findings: 23 May 2024. Four months later Steven Kramer was indicted in New Hampshire on 26 charges and the FCC proposed a $6 million fine; the cited sources do not tie those actions to Pindrop's analysis.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Assessed benefits: Society.
Depth: Society: 1.
Reasons: Society: Public attribution of the robocall's toolchain, corroborated by an academic team. No enforcement outcome in the sources is traced to the analysis.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: Independently reported

Kind: Not assessed
Related entries: Not assessed
Field: Election integrity / deepfake detection. Technology.
Purpose: Identify synthetic (deepfake) audio and attribute it to the generating system.
Dates: Date basis unassessed: 2024 (precision unassessed). Updated: 22 October 2026.
Stage: Not assessed
Version: SHA-256: 8214b8556bab…