Ethical AI & Responsible Data Collection Policy
Version 1.0 — effective 15 September 2026. This policy applies to all Posterior products, research and data services, and supplements our Company Privacy Policy and Data Privacy Policy.
At Posterior, we build AI systems and the data that trains them. This policy explains how we do both responsibly — with respect for the people who contribute, the communities our systems affect, and the law.
1. Purpose and scope
This policy covers all AI research, product development and data programmes at Posterior, including collection, annotation, review and delivery of AI training data. It applies to all staff, contributors, contractors and suppliers who act on behalf of Posterior.
2. Principles
We are guided by consent, dignity, fairness, transparency and accountability. Every programme must be lawfully collected, honestly produced and independently reviewable.
3. Responsible collection
We collect only what your brief requires and only for the purpose described. All collection is consent-first:
- Signed consent is obtained before any recording, with AI-training use disclosed. No recording starts without it, and consent status is verified before each session.
- All contributors are voluntary adults aged 18 or over. Eligibility is confirmed at onboarding.
- All content is scripted and fictional. We do not collect real private moments, and takes that break our content rules are rejected at capture, not cleaned downstream.
- Collection is purpose-limited to the approved brief. We do not repurpose data beyond the agreed programme without a new consent.
We do not scrape third-party content and we do not infer sensitive attributes about contributors.
4. Responsible annotation and review
Annotation is human-led. Machine drafts may accelerate work, but the human decision is the label. Every file is accepted or rejected by a reviewer who is independent of the collector, with reasons recorded. Raw files are never overwritten, and every delivery carries a per-file chain of custody.
5. Privacy and minimisation
We exclude personal data at capture rather than scrubbing it later. Framing rules, scripted content and a capture stack that avoids device identifiers, precise location and other unnecessary data keep most personal data out of the dataset by design. Contributor identities and signed forms remain in a private, access-controlled registry that never ships to clients. Deliveries carry only contributor and consent codes and per-file checksums.
Technical controls include raw immutability and a per-file chain of custody.
6. Regulatory alignment
Home-base practice is aligned with India’s Digital Personal Data Protection Act 2023. For other jurisdictions, safeguards — including GDPR-style data processing terms for EU/UK and CCPA handling for California — are set per client and on a case-to-case basis, proportionate to the data involved and recorded in the programme agreement. Programmes in regulated domains proceed only after a documented joint safeguards review.
7. Fair treatment
Contributors are treated fairly and with dignity. Participation is voluntary, roles and expectations are explained before work begins, pay is on the agreed terms and on time, and concerns may be raised without fear of retaliation. Coercion, forced labour and child labour are prohibited in our operations and supply chain.
8. Responsible AI development and use
We build models and products to be useful and safe. We consider intended use, limitations and potential harms during design, we evaluate performance before release, and we do not support uses that violate law or our content rules. Where a programme raises heightened ethical risk, it proceeds only after a joint review that is documented in writing.
9. Accountability and governance
This policy is owned by the Posterior leadership team and reviewed at least annually and after any incident. All persons to whom it applies are expected to read it, follow it and raise concerns promptly through integrity@posterior.xyz or to the team lead. Reports may be made anonymously, confidentiality is protected, and retaliation is prohibited.
Related documents: Company Privacy Policy, Data Privacy Policy, Terms of Service and Trust. Our Code of Conduct is maintained as a comprehensive internal policy and shared with clients per engagement.
Document history
- v1.0 — 15 September 2026 — Initial publication.
Contact for this policy: support@posterior.xyz — we respond within 30 days. For confidential reports: integrity@posterior.xyz.
