AI / Synthetic Consent Addendum (Creator Opt-In)
Creator: {{CREATOR_NAME}} Creator Account ID (internal): {{CREATOR_ID}} Effective Date: {{EFFECTIVE_DATE}} Asset(s) Covered: {{ASSET_LIST_OR_ASSET_CLASS}} Order Form(s) / Dataset(s): {{ORDER_FORM_IDS_OR_DATASET_IDS}}
This AI / Synthetic Consent Addendum is a separate opt-in agreement governing how my voice-derived materials may be used for AI, synthetic, or model-training purposes. It applies only to the checkbox selections I make below.
Identity & Authority (Required)
[ ] I CONFIRM that I am the voice owner and/or I have the legal authority to contribute and license the Voice Asset(s) covered by this agreement (including any necessary permissions from other rights-holders).
1) Definitions (Plain Language)
- Voice Asset(s) : recordings I upload or contribute to the platform.
- Derived Data : non-audio outputs computed from my Voice Asset(s), such as acoustic features, embeddings, labels, and QC metrics.
- Raw Audio : the original audio recordings (or watermarked versions) containing my voice waveform.
- Synthetic Use : training, tuning, evaluation, or operation of ML/AI systems that learn from my Voice Asset(s) or Derived Data.
2) Consent Scope (Check One or More)
A) Derived-Only Training & Analysis (Recommended / Lowest Risk)
[ ] I CONSENT to the use of Derived Data only from my Voice Asset(s) for ML/AI purposes, including:
- dataset analytics, feature extraction, embedding generation
- model training, fine-tuning, evaluation using derived representations
- distribution of Derived Training Packages to buyers (e.g., Parquet + manifest + QC + vocab)
What is NOT included: raw audio delivery to buyers, transcript text delivery unless separately opted-in.
Derived Data Not for Identification (Required): [ ] I ACKNOWLEDGE that Derived Data is intended for modeling utility and analytics and is not intended for identifying me or verifying my identity. I understand buyers are contractually prohibited from using Derived Data for identification, authentication, surveillance, or re-identification purposes.
Output / Access Limits:
- Max usage: {{MAX_USAGE}} (e.g., unlimited training use within license term or X hours per term )
- Term: {{TERM}}
- Buyer reach: {{REACH}} (e.g., single buyer org / multi-buyer catalog )
B) Optional: Transcript / Text Outputs (Higher Privacy Risk)
[ ] I CONSENT to share transcription text and/or segments derived from my audio as part of AI datasets.
Limits: {{TEXT_LIMITS}} Notes: If unchecked, transcription content is excluded or redacted in enterprise exports.
C) Optional: Raw Audio Access (Highest Biometric Risk)
[ ] I CONSENT to allow raw audio (or watermarked audio) to be delivered to buyers and/or used directly for model training.
Limits:
- Max minutes of delivered audio: {{MAX_AUDIO_MINUTES}}
- Term: {{TERM}}
- Distribution: {{RAW_AUDIO_DISTRIBUTION_SCOPE}} (e.g., single buyer only , restricted catalog )
If unchecked, raw audio is not licensed for buyer download/delivery under this addendum.
D) Optional: Synthetic Voice Generation / Personalization
[ ] I CONSENT to allow synthetic voice generation systems to be trained/tuned using my materials only within the limits above and only for permitted buyer uses.
Limits: {{SYNTH_LIMITS}} Important: This does not permit identity verification or impersonation outside the contract controls.
3) Prohibited Uses (Always Prohibited)
Even if I opt in above, my voice materials may NOT be used for:
- identity verification / authentication / speaker recognition intended to identify me
- surveillance, tracking, or law enforcement targeting
- hiring, employment decisions, eligibility, credit, housing, insurance, or admissions decisions
- medical diagnosis, mental health inference, disability inference, or health profiling
- generating content that is defamatory, harassing, sexually explicit involving minors, or otherwise illegal
- training systems intended to imitate me for fraud, deception, or impersonation
- using Derived Data (including embeddings) for speaker identification, authentication, re-identification, or linking my identity across datasets
4) Disclosure, Auditability, and Attestation
The platform may:
- apply watermarking or other attribution signals (where applicable)
- generate export manifests that record policy versions, artifact hashes, and QC results
- maintain logs required to demonstrate licensing compliance and payouts
Buyers may receive verification artifacts (e.g., manifest.json, QC reports) to validate dataset integrity.
5) Withdrawal / Revocation (Prospective)
I may withdraw this consent at {{WITHDRAWAL_METHOD}}.
- Withdrawal applies prospectively: it stops future licensing/delivery and future inclusion in new exports or refreshes after {{WITHDRAWAL_EFFECTIVE_WINDOW}}.
- I understand that previously delivered datasets or models trained prior to withdrawal may not be technically un-trained. Buyers remain bound by their contractual restrictions for future use.
6) Data Handling & Retention (BIPA-First)
- The platform will process my materials according to the published privacy policy and retention schedule: {{RETENTION_POLICY_REFERENCE}}.
- Enterprise exports may exclude direct identifiers and may be limited to Derived Data in strict modes.
- Where required by law, biometric or voice-derived processing will be supported by consent and documented controls.
7) Compensation Acknowledgment
If my materials are licensed for AI/Synthetic uses, compensation and royalty allocation will follow the applicable Order Form and Creator Terms (including Cohort Pro-Rata rules where specified): {{COMPENSATION_REFERENCE}}.
8) Acceptance
By selecting the consent options above and signing/accepting, I acknowledge:
- I had the opportunity to review what I am opting into
- my consent is limited to the checked options and stated limits
- prohibited uses remain prohibited regardless of consent selections
Creator Signature / Acceptance: ___________________________ Date: ___________________________