Who this applies to
People, partners & projects
Our team, eligible contributors, reviewers and prospective AI project partners. These are the standards Sprig adopts for the AI work it chooses to accept.
Responsible AI • Public policy v1.0 • 10 October 2026
AI work deserves responsible data practices.
Sprig Rewards publishes these standards to make our expectations for human-reviewed AI tasks, contributor rights, privacy and partner projects transparent.
Implementation status
Commitments, not certifications.
Sprig is early-stage. We describe our existing manual qualification controls separately from the additional data-governance steps required before accepting externally commissioned projects.
Our responsible AI standards
These commitments guide project screening and our development roadmap. They do not replace applicable laws, agreements, or the Sprig Privacy Policy.
1. Purpose, scope and ownership
This policy covers AI-related work that Sprig Rewards may organize, including text and image annotation, data collection, evaluation, contributor screening, quality review, storage, transfer and partner delivery. It applies to staff, reviewers and contributors when acting for Sprig. The platform operator is responsible for reviewing each project before acceptance.
2. Human oversight and honest representation
A human reviewer must be able to check deliverables and resolve uncertain or disputed judgments. Sprig will not market unverified contributor capacity, claim automatic AI-use detection, fabricate client relationships, misrepresent training data as human-created or promise earnings or work that is unavailable. Browser activity and webcam interruption indicators are review signals, never conclusive proof of misconduct.
3. Consent, lawful sourcing and data rights
Before collecting data for a partner, Sprig must identify the intended purpose, collection method, allowed uses and the party authorized to license the material. Projects must use lawfully collected or appropriately licensed data. Where people, voices, faces or sensitive information are involved, staff must confirm an appropriate legal basis and specific, informed contributor or subject consent where required. Scraping restricted content, reusing third-party copyrighted materials without rights, and deceptive collection are prohibited.
4. Privacy and data minimization
Only request data necessary for an approved task. Avoid collecting credentials, financial account secrets, precise location, children's data, health information, biometrics or other sensitive personal information unless a documented lawful basis, safeguards and project-specific authorization exist. Contributors must not upload personal information belonging to others without suitable authorization. Privacy requests are directed to contact@sprigrewards.com and handled under applicable rules and the published Privacy Policy.
5. Quality assurance and provenance
Each accepted project needs a documented task rubric, acceptance criteria, contributor eligibility and review process. We aim to log submission identifiers, task and contributor identifiers, time, review outcomes, source/rights evidence as applicable, and reviewer decisions. A reproducible asset-level chain of custody is required before a client project that demands full provenance is accepted. Sprig does not currently claim universal full-asset provenance or audited quality metrics.
6. Contributor fairness and safety
Disclose task instructions, eligibility, reward conditions and review requirements before participation. Provide a support route for disputes. Avoid coercive collection, unnecessary identification, dangerous or exploitative tasks and discriminatory treatment. A completed qualification is not an employment offer, promise of paid work or guarantee of acceptance. Do not require unnecessary exposure to disturbing content.
7. Security and access controls
Use access-limited systems for private submissions and interview recordings, signed upload or retrieval mechanisms when available, and administrative review permissions. Share personal data only with authorized people or processors for a documented purpose. Before accepting partner data, verify any client-specific encryption, access, location, confidentiality, security questionnaire and deletion requirements. Sprig does not claim ISO 27001, SOC 2 or other independent certifications.
8. Video qualification and retention
The current AI task qualification workflow requires explicit consent to webcam and microphone recording on desktop, asks ten spoken questions, and uses manual scoring with an 8/10 passing threshold. Recordings are stored in a private bucket for authorized reviewer access and a deletion attempt is performed after review. Storage deletion may fail and must be verified or retried; abandoned recordings are cleaned opportunistically by the current review workflow, not by a guaranteed time-based deletion service. Recording is not used as definitive evidence of AI assistance.
9. Client and project acceptance gate
Before publishing any paid AI project, Sprig must verify a legitimate authorized partner, permitted data rights, privacy terms, participant suitability, expected workload, review rubric, funding and payout terms, data transfer location and retention/deletion conditions. Projects that cannot meet these requirements must not be published or delivered. Partner requests are proposals until agreed in writing.
10. Restricted and unacceptable uses
Sprig will not knowingly support deceptive impersonation, non-consensual intimate imagery, illegal surveillance, unauthorized facial or biometric collection, credential theft, fraudulent datasets, abusive profiling, prohibited exploitation or other unlawful data uses. Potentially high-risk projects involving biometric, medical, child-related or sensitive demographic information must be declined or escalated for a qualified legal and safety review before collection.
11. Reporting, incidents and appeals
Contributors and partners should report suspected privacy violations, questionable rights, quality manipulation, security incidents or unsafe instructions to contact@sprigrewards.com. Sprig should restrict affected access or pause activity while an incident is assessed, preserve necessary review evidence, notify relevant parties where required and document the outcome. Qualification and reward decisions may be queried through support; reconsideration is not automatic.
12. Governance, updates and implementation
The operator reviews this policy when processes, partners or relevant laws change. Public commitments must be supported by actual controls. Project-level consent records, individual asset provenance, formal rights clearance, incident response procedures and jurisdiction-specific compliance remain subject to implementation and verification before relevant client work begins. Publication of this policy alone is not certification or legal compliance assurance.
Questions or concerns?
Report a data concern.
For a specific concern about data rights, AI task instructions or privacy, contact the Sprig operator. Do not attach sensitive client materials without an agreed secure transfer route.
Contact our team ↗Related resources
Read the complete context.
For how we handle personal information and participation eligibility, also read these documents.
Policy reference: SPRIG-RAI-001 · Version 1.0 · Published 10 October 2026 · Owner: Sprig Rewards platform operator · Review: as relevant regulations, service providers, project controls or data practices change. This document is not an independent certification or a legal opinion.