Back to AI Tools
Riverside.fm
Risk Overview
Moderate Risk
4.4
/10
Speech to Text
·
Audio Generation
·
Audio Editing
·
Top Risks
Mandatory binding arbitration combined with class-action waiver (score 7) eliminates court access and collective remedies. Standard-tier users' recordings are used for AI model training by default with no contractual opt-out mechanism (score 5), and data retention timelines remain unclear in the outdated privacy policy (score 5).
Mitigations
Negotiate a written Data Processing Addendum to opt out of training use and clarify retention commitments—Riverside has demonstrated willingness via its enterprise DPA (Dec 2024). Document all material terms in writing and engage legal review before signing; consider whether the business value justifies acceptance of binding arbitration, since the clause applies only to disputes arising from the ToS itself.
Get started
Is your team using
Riverside.fm
safely?
Highest Risk Categories
The highest-scoring risk categories identified in this assessment.
Legal Terms
ToS contains mandatory binding arbitration clause: 'any dispute, claim or controversy arising out of or relating to these Terms or your use of services shall be determined by binding arbitration on an individual basis rather than in court.' Class action waiver: 'Users and Riverside waive the right to a trial by jury or to participate in a class action.' Small claims court exception applies. This is the most aggressive legal posture found in this batch—mandatory arbitration plus class action waiver is the combination that triggers highest legal risk scores.
7
/10
Training Use
Privacy Policy (June 15, 2026) explicitly states Riverside processes Content Data as independent controller to train, develop, and improve its AI/ML models. Business customers with written DPA can opt out; standard/free tier has no contractual opt-out mechanism. ToS (July 2026) permits use of Inputs and Outputs to train models; prohibits users from training competing models but authorizes Riverside training. Compared to calibration: standard tier is closer to AIVA (score 6, biometric training) than Amazon Bedrock (score 1, no-training commitment) because default training applies to content without explicit per-use consent.
5
/10
Data Retention
Privacy policy (dated June 22, 2022—notably outdated) does not specify retention periods clearly. Recordings are stored in user accounts until deleted. No explicit auto-deletion or post-deletion window stated in accessible policy text. The DPA (Dec 2024 version) governs business customer data processing. Guest recordings retention depends on host account settings. Recording retention is a concern for a podcast tool where third-party guests participate.
5
/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://riverside.com/terms-conditions
Privacy Policy
https://riverside.com/privacy-policy
End User License Agreement (EULA)
https://riverside.fm/data-processing-addendum