Skip to main content
YPAI
Services Data Industries Company
AI Data & Evaluation
Data collection and sourcing Consent-led multimodal collection. Dataset licensing Rights-cleared datasets, ready to license. Data annotation Boxes, masks, spans and tracks, reviewed. Data labeling Class definitions, tie-break and the record. Model and agent evaluation Human evaluation and regression testing. Explore AI Data & Evaluation Create, source and evaluate the data your AI depends on.
AI Implementation
Discovery and architecture Scope the use case and the system design. AI assistants and chatbots Customer-facing conversation, service and voice. RAG and knowledge systems Retrieval over your own knowledge. Document AI and workflows Documents read, checked, routed and written once. Agents and workflow automation Agents and automation in production. Private and enterprise deployment Private, controlled deployment. Explore AI Implementation Turn a defined AI use case into a system you can operate.
Delivery
Connected Delivery Data, evaluation and implementation under one structure. Pilots Validate the delivery method before scale.
Explore all services
AI Data & Evaluation
Speech & Audio Data Multilingual speech, acoustic environments and voice data. Image Data Still images under specified coverage, device, rights and residency. Image, 3D & Sensor Data Point clouds, depth and multi-sensor rigs, calibrated and time-synchronised. Video Data On-camera and conversational video, collection through delivery. Physical AI Data Demonstration, episode and real-world physical data. Dataset Licensing & Sourcing Rights-cleared datasets, bespoke sourcing and acquisition. Data Annotation & Review Ontology design, marking, review and model-ready delivery. Data Labeling Class definitions, tie-break and the record of who decided. Model & Agent Evaluation Human evaluation, multilingual testing and failure analysis.
Explore AI Data & Evaluation
Operating conditions
AI Companies & Model Developers Training data, preference data and evaluation loops. Automotive & Mobility In-cabin speech, perception, video and sensor data. Financial Services Document AI, knowledge systems and traceability. Healthcare & Life Sciences Specialist data, domain review and privacy-sensitive work. Industrial & Energy Field data, operational workflows and integration. Public Sector Controlled data operations and reviewable AI systems.
Explore industry solutions
Company
About YPAI Company, mission, operating model and delivery history. Partnerships Commercial, technology and delivery collaboration. AI Blog Research, technical perspectives and company updates. Contact Projects, partnerships, procurement and general enquiries.
Become a Contributor Contact us
YPAI
Start here Scope a project Start small Pilots
AI Data & Evaluation
Collect
Data collection and sourcing Speech & Audio Data Image, 3D & Sensor Data Video Data Physical AI Data
Produce
Data Annotation & Review Data Labeling
Model & Agent Evaluation
Dataset Licensing & Sourcing
AI Implementation
Discovery and architecture
Build
AI assistants and chatbots RAG and knowledge systems Document AI and workflows Agents and workflow automation
Private and enterprise deployment
Connected Delivery
Operating conditions
AI Companies & Model Developers Automotive & Mobility Financial Services Healthcare & Life Sciences Industrial & Energy Public Sector
Also
Image Data
About YPAI Partnerships AI Blog

AI data, evaluation and implementation under one accountable delivery model.

Contact us Become a Contributor

Legal

Responsible Data Collection Policy

Last updated: 31 July 2026

This policy describes how YPAI collects, documents, reviews, stores and deletes data obtained from contributors for customer engagements. It applies to data collection activities conducted within AI Data & Evaluation and AI Implementation.

On this page

  • 1. Purpose and scope
  • 2. Principles
  • 3. How contributors are engaged
  • 4. Consent and lawful basis
  • 5. Provenance and traceability
  • 6. Data subject rights and withdrawal
  • 7. Storage, residency and transfers
  • 8. Quality and human review
  • 9. Documentation provided to customers
  • 10. Retention and deletion
  • 11. Boundaries and shared responsibility
  • 12. Governance of this policy

1. Purpose and scope

This policy describes how YPAI collects, documents, reviews, stores and deletes data obtained from contributors for customer engagements. It applies to data collection activities conducted within AI Data & Evaluation and AI Implementation. Controlled Delivery provides the shared operating framework for scope, rights, quality, security, change and acceptance from project definition to handover.

The policy applies to YPAI, contributors participating in a collection, and customers reviewing how collected data was obtained. Project-specific terms remain defined in the applicable project terms, data processing agreement, acceptance plan and transfer arrangements.

2. Principles

Purpose-specific collection. YPAI documents consent for each contributor and each stated purpose. Data rights and licence terms are linked to the purpose for which the data is collected.

Traceability. YPAI maintains a traceable connection between collected data, the contributor, the consent record and the applicable rights basis. Provenance records and dataset change logs preserve that connection through delivery.

Documented control. Collection, review and acceptance follow the scope and sampling plan agreed for the engagement. The resulting records allow the customer to examine how the dataset was assembled and reviewed.

Project-specific handling. Data residency, subprocessors and international transfer controls are defined for each project. European storage is the default position.

Withdrawal and deletion. YPAI maintains a documented withdrawal workflow and records the actions taken. At the end of a contract, YPAI applies its contractual 30-day erasure commitment.

3. How contributors are engaged

YPAI recruits contributors through its own channels rather than third-party marketplaces. Its contributor network includes 210,000+ contributors across 50+ countries.

Contributors participating in collection are identity-verified. YPAI uses its self-hosted annotation infrastructure for annotation work.

Contributor payments are made monthly by international transfer. Supported payment currencies are USD, EUR, NOK, BRL and SEK.

4. Consent and lawful basis

YPAI documents the lawful basis applicable to each collection purpose under GDPR Article 6. Where consent is the lawful basis, YPAI records consent per contributor and per purpose in accordance with Article 7. Where the collection involves special categories of personal data and Article 9 applies, YPAI documents explicit consent for the specified purpose.

Consent is not inferred from acceptance of general platform terms. The consent record is maintained separately and linked to the relevant purpose.

Each collected record is linked to its consent record through SHA-256 provenance manifests. This provides evidence of the connection between the record, the contributor, the stated purpose and the rights or licence basis associated with the dataset.

5. Provenance and traceability

YPAI maintains provenance information for collected records. For recorded data, this includes per-recording provenance linked to the contributor and the applicable consent record.

Rights and licence information remains traceable from the delivered dataset back to the consent on which collection and permitted use are based. Dataset versions are immutable and are accompanied by change logs so that changes between versions can be identified.

The provenance package is part of the documentation supplied for the engagement. It supports customer review of source, purpose, rights and dataset history.

6. Data subject rights and withdrawal

YPAI maintains a workflow for data subject rights and related restrictions under GDPR Articles 12 to 23, as applicable to the processing activity. The workflow records the request and the actions taken.

A contributor may withdraw consent through the withdrawal process. YPAI maintains an audit trail for withdrawal handling. The contractual 30-day erasure commitment at the end of a contract does not set a response period for an individual withdrawal request.

Where YPAI processes personal data on behalf of a customer, the respective responsibilities for handling requests are defined in the project terms and the applicable Article 28 data processing agreement.

7. Storage, residency and transfers

YPAI uses European storage by default. The agreed data residency position, authorised subprocessors and controls for international transfers are defined for each project.

YPAI provides standard Article 28 data processing agreement terms. Standard contractual clauses are available where the agreed processing or international transfer arrangement requires them. Article 28 controller-processor clauses and clauses for transfers outside the EEA address separate contractual functions.

Subprocessor information is included in the documentation supplied to the customer. The project terms define the processing arrangement that applies to the engagement.

8. Quality and human review

Human quality assurance is applied according to the acceptance and sampling plan agreed for each engagement. The plan defines where human review is used and how reviewed output is assessed for acceptance.

YPAI records QA artefacts and validation logs produced under that plan. These records document the review performed for the contracted scope without representing that every item receives the same review treatment.

Quality decisions, changes and acceptance are handled through Controlled Delivery and the terms agreed for the engagement.

9. Documentation provided to customers

For data collection engagements, YPAI provides documentation covering the controls applied to the dataset. The documentation includes, as applicable to the collected data:

  • per-recording provenance;
  • consent records per contributor and per purpose;
  • demographic and dialect distribution metadata;
  • QA artefacts and validation logs;
  • immutable dataset versioning with change logs;
  • subprocessor transparency;
  • sampling methodology documentation.

YPAI aligns this documentation with the data governance subjects addressed by EU AI Act Article 10 for high-risk AI systems. The documentation is structured so a conformity assessment can use it.

10. Retention and deletion

At the end of the contract, YPAI applies a contractual 30-day erasure commitment. The commitment concerns end-of-contract erasure and is separate from the workflow used for individual withdrawal requests.

Deletion is handled within the contracted project scope. The relevant project documents define the data, processing arrangement, subprocessors, residency position and transfer controls to which the commitment applies.

YPAI records withdrawal actions through the withdrawal audit trail and maintains dataset change information through version records and change logs.

11. Boundaries and shared responsibility

Where the customer acts as controller, it retains responsibility for the purposes for which it uses the data, the instructions it gives to YPAI, the intended use of its AI system and its own conformity assessment. A processor acts on behalf of the controller and according to the controller's documented instructions.

YPAI is responsible for applying the controls stated in this policy within the agreed scope. YPAI provides the provenance, consent, distribution, quality, versioning, subprocessor and sampling documentation described above to support customer review.

Project terms, the data processing agreement, standard contractual clauses where applicable, and the acceptance plan govern the engagement alongside this policy.

12. Governance of this policy

This policy is reviewed according to YPAI's assigned review cadence and when material changes to services, evaluation methods, regulatory posture, subprocessors, residency controls or monitoring practices affect its content.

Questions about this policy are raised through the YPAI contact used for procurement, risk, security or the relevant engagement. YPAI records the question, provides a documented response and updates this policy where a change is required.

Version: 1.0

Effective date: 31 July 2026

Document owner: Data Protection Officer

Questions about this policy can be raised through the contact form. The governance this policy operates under is set out in the Ethical AI Policy.

Start with the system or the data.

YPAI builds production AI systems and delivers the multimodal data used to train, evaluate and improve them.

Contact us Scope a pilot

AI systems, data and evaluation under one accountable delivery model.

New projects · accepting data and AI work
Engagement scoped before build
Acceptance defined before delivery
Services
AI Data & Evaluation AI Implementation Controlled Delivery Dataset Licensing
Capabilities
Speech & Audio Image Data Image, 3D & Sensor Data Video Data Video Data Collection Physical AI Data Annotation & Evaluation Data Labeling
Company
About YPAI Partnerships Contact Become a Contributor
Resources & Legal
AI Blog Privacy Contributor Terms Cookie Policy Data processing
YPAI · Oslo, Norway · Global delivery
Disclaimer LinkedIn ↗
EEA RESIDENCY BY DEFAULT · ARTICLE 28 DPA TERMS AVAILABLE
© 2026 YPAI