Responsible AI

1. Purpose and scope

This Responsible AI statement applies to AI-enabled features, automated analysis, generated recommendations, mappings, rules, documentation, code, workflows and other outputs provided through Singularity Engine.

It is intended to complement our Privacy Policy, Terms of Use, customer agreements, security practices and applicable laws.

2. Our Responsible AI principles

3. Human oversight and decision control

Singularity Engine is built around an assisted-decision model rather than unrestricted autonomous execution. Depending on the engagement, AI may propose classifications, mappings, rules, architecture patterns, transformations or implementation steps.

Appropriate human reviewers should:

• Review material recommendations before approval or implementation.
• Validate generated code, rules, mappings and architecture decisions.
• Apply domain, regulatory, security and operational judgement.
• Define approval thresholds based on risk, impact and confidence.
• Retain the ability to reject, modify, pause or reverse AI-supported actions.

4. Data responsibility and privacy

We seek to use only the information required for the authorised purpose. Customers should avoid providing personal, confidential, regulated or sensitive data unless it is necessary and covered by appropriate contractual and technical controls.

We do not use customer data to train public or shared AI models unless the customer has provided explicit written authorisation.

Our approach may include:

• Data minimisation and purpose limitation.
• Role-based access controls and environment separation.
• Masking, redaction or tokenisation where appropriate.
• Defined retention and deletion practices.
• Evaluation of third-party model and service-provider data practices.

5. Transparency and explainability

We aim to clearly distinguish AI-generated or AI-assisted outputs from independently verified facts. Where practical, Singularity Engine may provide supporting context such as source references, confidence indications, assumptions, lineage, rules, decision logs or validation status.

The level of explanation required should reflect the risk and impact of the use case. Higher-impact decisions require stronger evidence, documentation and human review.

6. Fairness, bias and inclusion

AI systems can reflect limitations or biases in data, models, prompts, assumptions and human decisions. Where relevant to the use case, customers and Singularity Engine should assess whether outputs could create unjustified disadvantage, exclusion or disparate impact.

Mitigation measures may include:

• Reviewing data sources, coverage and representativeness.
• Testing outputs across relevant scenarios and user groups.
• Using human escalation for uncertain or high-impact results.
• Documenting known limitations and residual risks.
• Monitoring material changes in model or output behaviour.

7. Security and resilience

We apply reasonable administrative, technical and organisational safeguards to protect AI-enabled services and associated data. Measures may include authentication, access control, encryption, logging, monitoring, secure development practices, vulnerability management and incident-response procedures.

Customers are responsible for securing their own environments, identities, credentials, integrations and downstream deployments.

8. Quality, validation and monitoring

AI outputs may contain errors, omissions, unsupported assumptions or unsuitable recommendations. Before relying on an output, users should assess its accuracy, completeness, relevance and fitness for the intended purpose.

Validation may include:

• Source-to-target reconciliation and evidence checks.
• Business-rule and data-quality validation.
• Code review, testing and controlled deployment.
• Comparison against approved architecture and governance standards.
• Monitoring for drift, degradation, anomalies or repeated failure patterns.

9. Customer responsibilities

Customers and authorised users are responsible for:

• Selecting suitable use cases and defining risk-based controls.
• Providing accurate, lawful and appropriately authorised data.
• Reviewing and validating outputs before business or production use.
• Maintaining required approvals, records and regulatory assessments.
• Ensuring qualified human oversight for material decisions.
• Monitoring deployed solutions and reporting suspected issues.

10. Prohibited and restricted uses

Singularity Engine must not be used to:

• Violate law, privacy rights, intellectual-property rights or contractual obligations.
• Facilitate discrimination, unlawful surveillance, deception or harmful manipulation.
• Generate or deploy malicious code or compromise system security.
• Make high-impact decisions about individuals without legally required safeguards and meaningful human review.
• Misrepresent AI-generated content as independently verified or professionally certified.
• Bypass safety, security, approval or governance controls.

11. Third-party AI models and services

Singularity Engine may use or integrate with third-party AI models, cloud platforms, data services or software components. These providers may have their own terms, limitations, availability, security controls and data-handling practices.

We assess third-party services in proportion to their role and risk, but we cannot guarantee that third-party models will always be accurate, available, unbiased or error-free.

12. Governance and review

Responsible AI is an ongoing process. We may review AI-enabled capabilities using risk assessments, architecture reviews, security controls, testing, approval workflows, documentation and monitoring.

Controls may vary based on the sensitivity of the data, level of automation, customer environment, regulatory context and potential impact of an incorrect output.

13. Reporting concerns and incidents

Users should promptly report suspected security issues, harmful outputs, privacy concerns, significant inaccuracies, misuse or other responsible-AI concerns. We will review reports and take proportionate action, which may include investigation, correction, restriction, suspension or control enhancement.

14. Changes to this statement

We may update this Responsible AI statement as our technology, services, risk controls or legal obligations evolve. The revised version will be posted on this page with an updated effective date.

15. Contact us

For questions or concerns about responsible AI at Singularity Engine, contact: