Insured international shippingGSG custody included with PremiumEscrow guaranteed on every transaction
Gold€123.85/ g-0.10%Silver€2.04/ g-0.04%Platinum€52.23/ g+0.15%
Skip to content
Gold Silver Global
Compare list empty
Back to legal index
Legal document

AI Usage Policy

How we use AI under Regulation (EU) 2024/1689 (AI Act), GDPR art. 22, the NIST AI RMF and US Executive Order 14110.

Updated: 2026-06-01

⚠️ Informational template. This document is a preliminary version pending review by counsel in each applicable jurisdiction (US + EU/Spain). It does not constitute binding legal advice.

Data controller and service operator:

1. Legal framework

  • Regulation (EU) 2024/1689 (AI Act) — risk-based classification, transparency, governance.
  • Regulation (EU) 2016/679 (GDPR) — art. 22 automated decisions.
  • LOPDGDD — digital rights (Spain).
  • NIST AI Risk Management Framework (AI RMF 1.0) — US reference.
  • Executive Order 14110 — Safe, Secure, and Trustworthy Development and Use of AI (US).
  • CCPA/CPRA — automated decision-making opt-out (California).

2. Guiding principle

AI within GSG never makes binding decisions. Its role is to suggest, translate, summarize and alert. Any decision on KYC, blocks, dispute resolution or escrow release is always human.

3. AI Act classification

Current GSG AI uses fall primarily within the limited risk band of the AI Act (transparency obligations). We do not run high-risk systems per Annex III (HR, credit, justice) in production. Any future feature falling into high-risk will be assessed against AI Act arts. 8-15.

4. Active AI functions

  • Listing moderation — pre-classification before human review.
  • Article translation — ES → EN/FR drafts subject to manual review.
  • Dispute summary — moderator suggestion.
  • Anti-fraud — suspicious pattern detection.
  • Dynamic spot pricing — optional alignment to live spot.

5. Transparency

Per AI Act art. 50 and the DSA:

  • AI-generated outputs are labeled (visible tag on translations and AI summaries).
  • Users know when they interact with an automated system.

6. Providers

GSG operates a proprietary AI gateway that abstracts the provider. API keys are stored encrypted with AES-256-GCM and never exposed to the frontend. Supported providers may include Ollama (local), OpenAI, Anthropic, Mistral and Google Gemini.

7. Data sent to AI

  • Only the minimum necessary for the function.
  • No KYC documents or banking data are sent.
  • Images may be watermarked before sending.
  • We comply with GDPR arts. 5 and 25 (minimization and privacy by design).

8. Logging and audit

Every AI call is logged with function, provider, model, latency, cost and outcome. Admin → Settings → AI displays the latest events, feeding our internal AI RMF aligned with NIST.

9. Your rights

For users in the EU / Spain

Per GDPR art. 22, you have the right not to be subject to solely automated decisions producing significant effects. Since GSG does not make such decisions solely automatically, this right is respected by design. You may also object to legitimate-interest processing (art. 21).

For users in California (CCPA/CPRA)

You may opt out of automated decision-making technology (ADMT) under upcoming CPRA regulations.

For users in the US generally

We honor reasonable requests under NIST AI RMF, Executive Order 14110 and FTC guidelines on AI.

10. Bias and evaluation

GSG conducts periodic reviews of deployed models to detect bias, errors and drift. Serious incidents are logged and, where applicable, reported under the AI Act.

11. Contact

dpo@goldsilverglobal.net. Last updated: 2026-06-01.