COMMERCIAL AI TRAINING AND FINE-TUNING DATA LICENSE

AI TRAINING AND FINE TUNING DATA LICENSE – COMMERCIAL USE, VERSION 1.0

Version 1.0

For production AI systems, SaaS, APIs, and monetised outputs.

Platform-Specific License Notice: > This License is issued by Opendatabay (www.opendatabay.com) to support both Data Suppliers and Data Buyers. Its use outside the Opendatabay platform, or without Opendatabay’s knowledge and authorization, is strictly prohibited. This License applies to Data Products listed on Opendatabay or provided via the platform, including samples, previews, or full datasets.

1. Definitions

"Data Product": The collection of bundled and packaged data being licensed, which may include text, audio, images, video, raw or structured data, as described in the accompanying documentation, together with any provenance, transparency, or compliance information provided by the Licensor to support AI training, evaluation, and model documentation (e.g., model cards or safety assessments).

"Licensee": The individual or entity receiving rights to use the Data Product under this License.

"Licensor / Data Provider": The individual or entity that collected, prepared, and owns the Data Product and is granting this License.

"AI / ML / LLM": Artificial Intelligence, Machine Learning, and Large Language Models, including neural networks or other model architectures trained or fine-tuned using the Data Product.

"Models": Machine learning models created, trained, or fine-tuned using the Data Product.

"Derivative Works": Outputs of Models, including model weights, embeddings, fine-tuned versions, or other representations derived from the Data Product.

"Outputs": Model predictions, generated text, audio, embeddings, or other results produced using the Data Product.

"Permitted Uses": Activities expressly allowed under this License, including Training, Fine-Tuning, Evaluation, Commercial Use, and creation of Derivative Works.

"Order": The purchase agreement, invoice, or transaction record that specifies the Data Product being licensed, pricing, payment terms, Effective Date, and contact information for both parties.

"Effective Date": The date specified in the Order when this License becomes effective, typically the date of purchase or payment confirmation.

2. License Grant and Scope (Permitted Uses)

2.1 License Grant

The Licensor grants the Licensee a non-exclusive, worldwide, perpetual (subject to termination rights in Section 10) right to use this Data Product solely for training, fine-tuning, evaluating, and developing AI, ML, LLMs, and agentic AI systems.

The Licensee may:

• Train or fine-tune models, including LLMs and AI agents, AI applications and systems.

• Create and use model outputs, embeddings, weights, or other Derivative Works.

• Use resulting models, systems, applications and outputs commercially.

2.2 Licensor Representations

The Licensor represents and warrants that:

• It has the rights and authority necessary to license the Data Product for the Permitted Uses.

• The Data Product was collected and provided in compliance with applicable data protection and related laws (including, where applicable, GDPR, EU AI Act and similar regimes).

• The Licensor has used commercial best practices and reasonable technical measures to filter and exclude personal data/PII or special category sensitive data from the Data Product.

• The Licensor is not aware of any malware, viruses, or malicious code in the Data Product.

• The Data Product has been prepared with reasonable care and professional standards for AI training purposes.

• To the best of Licensor’s knowledge, the Data Product has not been intentionally altered to include ‘poisoned’ samples, trigger phrases, or adversarial inputs designed to compromise the alignment or security of resulting Models.

The Licensor does not warrant that the Data Product is free from all third-party claims, nor that Licensee’s specific use of the Data Product, Models, or Derivative Works will be lawful in all jurisdictions or for all intended use cases. The Licensor does not guarantee specific model performance or outcomes.

3. Restrictions (Prohibited Uses)

This license covers the following activities only: training, fine-tuning, evaluating, and developing AI, ML, LLM, and agentic AI systems.

3.1 Prohibited Activities

Licensee may not:

• Resell, redistribute, or sublicense the Data Product itself (see Section 4 for sublicensing of Models and Derivative Works).

• Use the Data Product for non-AI purposes such as business analytics, market research, direct data distribution, or database compilation.

• Intentionally remove, obscure, or alter any digital watermarks, cryptographic signatures, or provenance metadata embedded within the Data Product. These markers are essential for verifying ethical sourcing and supply chain integrity.

• Use the Data Product in ways that violate applicable laws or regulations, including discrimination, unlawful surveillance, or other prohibited applications.

3.2 Separate Licensing for Other Uses

For uses outside AI training/fine-tuning (e.g., analytics, visualization, reporting, or general data services), contact Licensor for separate licensing terms.

4. Models, Derivatives, and Redistribution Rights

4.1 Rights Granted

Licensee may create Models using the Data Product, including training, fine-tuning, or evaluating AI, ML, LLMs, and agentic AI systems. Licensee is granted a non-exclusive, worldwide, and perpetual right to:

• Use Model outputs (predictions, text, audio, embeddings, images, or other results) for any lawful purpose consistent with this License, including commercial use.

• Create and use Derivative Works such as fine-tuned weights, embeddings, or model variants.

• Evaluate Models and publish benchmarks or performance results, provided that the Data Product itself is not disclosed.

4.2 Restrictions on Data Product Redistribution

Licensee may not redistribute or sublicense the original Data Product itself unless explicitly granted a separate license from the Licensor.

4.3 Sublicensing of Models and Derivative Works

Licensee may sublicense or provide access to Models and Derivative Works created using the Data Product (including fine-tuned weights, embeddings, and AI-based services), provided that:

• The underlying Data Product is not made available, exposed, or reasonably reconstructable.

• All downstream recipients are informed that they receive no rights to the Data Product beyond those expressly granted in this License.

• Any sharing of Derivative Works must not expose, reconstruct, or reveal the underlying Data Product.

Safe Harbor for Model Outputs: The inadvertent, non-systematic reproduction of insignificant portions of the Data Product by a Model (e.g., through overfitting or stochastic generation) shall not constitute a breach of this Section, provided that the Licensee has not intentionally engineered the Model to act as a database or substitute for the Data Product.

4.4 Responsibilities and Liability

Licensee is responsible for its own use of Models and Derivative Works and any direct downstream recipients to whom it sublicenses. Licensor acknowledges that once Models are released (particularly as open weights), Licensee cannot control all downstream usage and is not liable for unforeseeable third-party misuse beyond its reasonable control.

While Licensee may release Derivative Works (such as Model Weights) commercially, Licensee acknowledges they are solely responsible for implementing ‘defence-in-depth’ safeguards and ensuring such releases comply with applicable Frontier AI Safety Frameworks. Licensor is not liable for downstream misuse of open-weight models derived from the Data Product.

5. Compliance Obligations

5.1 Legal and Regulatory Compliance

Licensee must comply with all applicable laws, including:

• GDPR and other data protection regulations

• EU AI Act transparency and documentation requirements, including Article 53 training data summaries where applicable

• Export control regulations (see Section 5.2)

• Copyright and intellectual property laws

• Anti-discrimination and human rights laws

5.2 Export Control and Sanctions Compliance

Licensee represents that it is not subject to trade sanctions or export restrictions under applicable law. Licensee will not use the Data Product or Models in violation of export control laws including U.S. Export Administration Regulations (EAR), International Traffic in Arms Regulations (ITAR), or equivalent regulations. Licensee acknowledges that AI models may generate outputs subject to export controls and is responsible for implementing appropriate compliance measures, including Technology Control Plans where foreign nationals access internal models.

5.3 Data Security and Protection

Licensee must implement appropriate technical and organizational measures to protect the Data Product from unauthorized access, disclosure, or misuse.

5.4 High-Risk AI Systems

For high-risk AI systems under EU AI Act or similar regulations, Licensee must maintain appropriate audit trails, bias assessments, and risk management documentation.

5.5 EU AI Act Transparency Requirements

For general-purpose AI models covered by EU AI Act Article 53, Licensee acknowledges responsibility for publishing training data summaries using the official EU Commission template and respecting copyright reservations as required by law. Licensor will provide reasonable cooperation for such disclosures, subject to protection of trade secrets and confidential information.

6. Data Retention, Deletion, and Audit

6.1 Retention

Licensee may keep copies of the Data Product only as long as needed for training, evaluation, or development of Models, and in line with applicable laws.

6.2 Deletion

Upon termination of this License (subject to Section 10.2) or written request from the Licensor, Licensee must promptly delete or irreversibly anonymize all copies of the Data Product, except for copies retained in standard backup systems that are scheduled for automatic deletion within a reasonable timeframe.

6.3 Models & Derivatives

Licensee may retain Models, Derivative Works, logs, and metadata indefinitely, provided the Data Product itself cannot be reconstructed from these materials.

6.4 Audit Rights

Where Licensor has reasonable suspicion of material breach of this License, Licensor may request reasonable evidence of compliance (e.g., deletion records, data retention policies, or security measures) no more than once per year, with sixty (60) days’ advance written notice. On-site audits, if needed and mutually agreed, are conducted at Licensor’s expense unless a material breach is discovered, in which case Licensee bears the costs. Audits are limited to deletion/retention records and compliance policies. Licensor shall have no access to proprietary Model architectures, business strategies, or other confidential business information beyond what is necessary to verify compliance with this License.

7. Intellectual Property, Attribution, and Trademarks

7.1 Ownership of Data Product

The Data Product, including all content, metadata, and structure, is owned by the Licensor. Except for the rights granted under this License, no ownership or interest in the Data Product is transferred to the Licensee.

7.2 Ownership of Models and Derivative Works

Licensee owns Models and Derivative Works it lawfully creates using the Data Product, but this ownership does not include or extend to the Data Product itself.

7.3 Attribution

Neither party may publicly reference the other party or disclose the other party’s use of the Data Product or this licensing relationship without explicit written consent, except as required by law or regulation. Licensee may choose to provide credit to the Licensor in public-facing materials (such as model cards or documentation), but is under no obligation to do so. No attribution should imply endorsement or sponsorship by Licensor without a separate written agreement.

7.4 Trademarks and Branding

All Licensor trademarks, logos, and trade names remain the property of the Licensor. Licensee may not use them in ways that imply endorsement, sponsorship, or partnership without separate written permission.

7.5 Third-Party Rights

Licensee is responsible for ensuring its use of Models or Derivative Works does not infringe third-party rights. This License covers only the Data Product; it does not grant rights to any third-party content included or referenced in the Data Product.

8. Warranty Disclaimer and Liability Limits

8.1 Limited Data and IP Warranty

Licensor warrants that:

• It has the right to license the Data Product under this License.

• The Data Product was not obtained in violation of applicable law or third-party intellectual property rights.

• The Data Product has been prepared with reasonable care and professional standards for AI training purposes. Licensor does not warrant that the Data Product is free from all third-party claims, nor that Licensee’s specific use of the Data Product, Models, or Derivative Works will be lawful in all jurisdictions or for all intended use cases. Licensor does not guarantee specific model performance or outcomes.

8.2 No Other Warranties

Except for the limited warranties set out in Section 8.1, the Data Product, and any related documentation, Models, or services, are provided “as is” and “as available.” To the fullest extent permitted by law, Licensor disclaims all other warranties, whether express or implied, including implied warranties of merchantability, fitness for a particular purpose, accuracy, completeness, non-infringement, or that the Data Product or Models will be error-free, uninterrupted, or suitable for any specific use.

8.3 No Guarantee of Results

Licensor does not guarantee any specific outcomes, performance, or results arising from the use of the Data Product, including any Models or Derivative Works created using it. Licensee is responsible for assessing and validating the suitability, accuracy, and safety of the Data Product, Models, and outputs for its intended use, including compliance with applicable laws, regulations, and industry standards.

8.4 Limitation of Liability

To the fullest extent permitted by applicable law, each party’s total aggregate liability to the other for all claims arising out of or relating to this License, the Data Product, or any related Models or services—whether in contract, tort, or otherwise—will not exceed the greater of: (a) the total amounts paid by Licensee for the Data Product under this License; or (b) any fixed liability cap specified in the Order or applicable commercial terms. This limitation applies in the aggregate to all claims and causes of action.

8.5 Exclusion of Certain Damages

To the fullest extent permitted by law, neither party will be liable to the other for any indirect, incidental, consequential, special, punitive, or exemplary damages, including loss of profits, revenue, goodwill, data, or business interruption, even if such damages were foreseeable or the party was advised of their possibility.

8.6 Carve-Outs and Mandatory Legal Exceptions

Nothing in this Section 8 limits or excludes liability that cannot be limited or excluded under applicable law, including liability for death or personal injury caused by negligence, or for fraud or fraudulent misrepresentation. The limitations and exclusions in Sections 8.4 and 8.5 do not apply to: (a) Licensee’s payment obligations; or (b) Either party’s indemnification obligations under Section 9, if applicable. Where liability cannot be excluded, the remaining limitations in this Section apply to the maximum extent permitted by law.

9. Indemnification

9.1 Licensor Indemnification

Licensor agrees to indemnify, defend, and hold harmless Licensee from and against any third-party claims, damages, losses, or expenses (including reasonable attorneys’ fees) arising from Licensor’s breach of the warranties in Section 8.1, provided that:

• Licensee promptly notifies Licensor in writing of the claim.

• Licensor has sole control of the defense and settlement of the claim.

• Licensee reasonably cooperates in the defense at Licensor’s expense.

This indemnification obligation does not apply to claims arising from Licensee’s modification of the Data Product, combination with other materials, or use outside the scope of Permitted Uses.

9.2 Licensee Indemnification

Licensee agrees to indemnify, defend, and hold harmless Licensor from and against any third-party claims, damages, losses, or expenses (including reasonable attorneys’ fees) arising from:

• Licensee’s use of the Data Product, Models, or Derivative Works in violation of this License.

• Licensee’s violation of applicable laws or regulations.

• Claims that Models or Derivative Works created by Licensee infringe third-party rights, except to the extent caused by the Data Product itself. Licensor must promptly notify Licensee in writing of the claim, allow Licensee sole control of the defense and settlement, and reasonably cooperate at Licensee’s expense.

10. Term, Termination, and General Terms

10.1 Term & Termination

This License starts on the Effective Date and continues perpetually, subject to the following termination rights:

• Either party may terminate for material breach if the breaching party fails to cure the breach within thirty (30) days of written notice.

• Licensor may terminate immediately if required by applicable law or court order, or if Licensee’s use creates substantial legal, security, or compliance risks that cannot be reasonably mitigated.

10.2 Effect of Termination

For termination due to Licensee’s material breach of payment obligations:

• All rights to use the raw Data Product end immediately.

• Licensee must delete the Data Product as per Section 6.2.

• Licensee may retain and continue to use Models and Derivative Works created prior to termination, provided that Licensee cures the payment breach (including any applicable interest) within thirty (30) days of termination.

For all other termination scenarios:

• All rights to use the Data Product end immediately.

• Deletion/anonymization obligations in Section 6.2 apply.

• Licensee may continue to use Models and Derivative Works created before termination under the terms of this License.

• Sections 4.4, 7, 8, 9, and 10 survive termination.

10.3 Governing Law & Jurisdiction

This License is governed by the laws of England and Wales, without regard to conflict of law principles. Any disputes arising out of or relating to this License shall be resolved exclusively in the courts of London, England.

10.4 Notices

All notices under this License must be in writing and sent to the contact information specified in the Order. Notices may be delivered by email (with read receipt or confirmation), registered mail, or courier service. Notices sent by email are deemed received upon confirmation of delivery. Notices sent by mail or courier are deemed received three (3) business days after sending (for domestic delivery) or seven (7) business days after sending (for international delivery). Each party is responsible for keeping its contact information current and notifying the other party promptly of any changes.

10.5 Assignment

Licensee may not assign, transfer, or delegate its rights or obligations under this License without Licensor’s prior written consent, which shall not be unreasonably withheld. Licensor may assign this License to affiliates or in connection with a merger, acquisition, corporate reorganization, or sale of substantially all assets related to the Data Product, provided the assignee agrees to be bound by this License. Any attempted assignment in violation of this section is void.

10.6 Entire Agreement and Amendments

This License, together with the Order, constitutes the entire agreement between the parties regarding the Data Product and supersedes all prior agreements, understandings, and communications, whether written or oral. This License may only be amended by a written document signed by both parties. No other modifications, including terms in purchase orders or other business forms, shall have any effect.

10.7 Severability

If any provision of this License is held to be invalid, illegal, or unenforceable, the remaining provisions shall continue in full force and effect. The invalid provision shall be reformed to the minimum extent necessary to make it valid and enforceable while preserving the parties’ original intent.

10.8 Waiver

No failure or delay by either party in exercising any right under this License constitutes a waiver of that right. No waiver shall be effective unless made in writing and signed by the party granting the waiver.

10.9 Platform Third-Party Beneficiary

The parties acknowledge that this License is facilitated by Opendatabay. The parties agree that Opendatabay is a third-party beneficiary of this License solely for the purposes of enforcing the prohibition on off-platform transactions or circumvention of Platform fees (if applicable under Opendatabay’s Terms of Service).

END OF LICENSE

This license provides balanced protection for both data sellers and buyers, with clear definitions, fair termination provisions, symmetric confidentiality, proportional liability, and compliance with current AI regulations including EU AI Act requirements and export control considerations.