Back to AI Tools
Resemble AI
Risk Overview
Moderate Risk
3.6
/10
Audio Generation
·
Voice Cloning
·
Top Risks
Voice cloning of real individuals' voices is inherently high-risk biometric data processing. Consent management responsibility falls on the customer — Resemble enforces consent at platform level but cannot verify real-world consent chain. Litigation environment around AI voice cloning is active (Lehrman & Sage case in same sector). Retention timelines for cloud-hosted voice models not fully specified.
Mitigations
Use on-premise or air-gapped deployment for maximum data sovereignty. Maintain rigorous written consent records for all voice cloning subjects. Use built-in audio watermarking for all outputs. Enable deepfake detection for output verification. Review and sign enterprise DPA for GDPR workflows.
Get started
Is your team using
Resemble AI
safely?
Highest Risk Categories
The highest-scoring risk categories identified in this assessment.
Legal Terms
Terms of service contain standard governing law provisions (California). Users must warrant they own or have rights to all audio submitted. No prominent mandatory arbitration or class action waiver identified in public terms. Company faced the Lehrman & Sage v. Lovo (related lawsuit in voice AI space, distinct company) indicating active litigation environment. Standard limitation of liability language.
5
/10
Training Use
Client audio data is used exclusively for training that client's AI voice model — not shared across customers. Biometric voice data is anonymized and securely stored per their ethics policy. Resemblyzer open-source speaker verification ensures consent matching. No cross-customer model training from individual voice clones. On-premise deployment available which keeps data off Resemble servers entirely.
4
/10
Data Retention
No specific public retention period stated for voice model data in standard hosting. On-premise deployment gives customers full control over retention. Cloud-hosted voice models persist for service continuity. Customers can delete voice models but exact post-deletion data scrubbing timelines are not publicly specified.
4
/10
Unlock the complete assessment
Run a free assessment to access the full AI tool risk analysis.
Updated July 2026 • AI-assisted research • Reviewed by Verax • Report an issue
Reference Documentation
Official documentation and policies referenced as part of this assessment.
Terms of Service
https://www.resemble.ai/terms-of-service
Privacy Policy
https://www.resemble.ai/privacy-policy/