Qstamp / Responsible SI Policy
SI governance

Responsible SI Policy

How Qstamp supports the accountable development and operation of SI systems, and the principles that govern its use.

1. Terms used in this policy

In its own wording, Quantova Inc uses the term SI, short for superintelligence, for advanced artificial intelligence systems and agents. Where a law, regulation or standard uses the term artificial intelligence or AI, such as the EU AI Act, this policy keeps the term used in that instrument. References in this policy to SI systems include AI systems as defined in that legislation.

2. Purpose

Qstamp is record keeping infrastructure for SI systems. It does not make decisions, generate content or process records on behalf of its users. Its purpose is to make the actions and artefacts of SI systems traceable and verifiable, so that operators can be held to account.

3. Principles

  • Accountability. Every anchored record is bound to the account that anchored it.
  • Traceability. Actions, models, datasets and policies can be linked to verifiable evidence of when they existed.
  • Transparency. The construction, verification procedure and source code are published.
  • Human oversight. Receipts give reviewers the evidence needed to examine and challenge automated decisions.
  • Integrity. Evidence remains verifiable through the transition to post quantum cryptography.

4. Alignment with the EU AI Act

Regulation (EU) 2024/1689, the EU AI Act, requires high risk AI systems to allow the automatic recording of events under Article 12, requires providers to keep documentation and automatically generated logs under Articles 18 and 19, and requires deployers to keep logs under Article 26. Article 50 sets transparency obligations, including the marking of content generated by AI systems.

Qstamp is designed to support these duties by making logs, documentation and records of generated content tamper evident and verifiable for their full retention period. Qstamp does not by itself classify an AI system, fulfil any obligation in full or establish conformity, and operators remain responsible for their own compliance.

5. Alignment with other frameworks

Qstamp supplies traceability evidence for the Govern, Map, Measure and Manage functions of the NIST AI Risk Management Framework, and for the record, monitoring and documentation controls of ISO/IEC 42001. Qstamp records can also support the documentation and traceability expected under other instruments, such as the Model AI Governance Framework published in Singapore and the Korean AI Framework Act. Quantova Inc does not claim certification under any framework, and Qstamp does not by itself establish compliance with any law.

6. Use of customer data

Quantova Inc does not receive the records that users stamp, and it does not use any customer record, fingerprint or receipt to train or evaluate any model.

7. The Ask Qstamp assistant

The Ask Qstamp assistant on this website is not an SI system. It selects its replies from a fixed set of answers written by Quantova Inc and contained in the page, and it sends nothing to any server.

8. Declared profiles

Operators may declare profile information about their SI systems, which QVMScan displays together with public network records on its Superintelligence Fingerprints page. A declared profile shows what the operator has declared. It is not a certification, endorsement or verification of the system or its operator by Quantova Inc.

9. Prohibited uses

Qstamp must not be used to lend false credibility to fabricated evidence, to conceal unlawful automated decisions, or in support of any practice prohibited under Article 5 of the EU AI Act. The Acceptable Use Policy sets out further restrictions.

10. Limits of evidence

A Qstamp receipt shows that a record existed in an exact form no later than a stated time. It does not show that the record is accurate, lawful or fair, and it does not establish the identity of a natural or legal person.

11. Raising concerns

Concerns about the use of Qstamp in connection with an SI system can be sent to [email protected].

12. Review

We intend to review this policy at least once a year and whenever relevant law or guidance changes.