Responsible AI

1. Our Commitment

Nubio Deep Intelligence Pte. Ltd. ("Nubio") builds AI systems that augment human decision-making, not replace it. We are committed to developing and deploying AI that is transparent, accountable, and aligned with the following frameworks:

• Singapore's Model AI Governance Framework (IMDA, updated January 2026), including the Agentic AI supplement.
• Applicable data protection obligations under the Personal Data Protection Act 2012 ("PDPA").

2. How Our AI Works

Our Services employ causal modelling — AI systems that learn the cause-and-effect relationships between operational variables, rather than relying solely on statistical pattern matching.

• Outputs are recommendations backed by causal reasoning. Users can understand why a recommendation was made, not just what was recommended.
• AI-generated outputs are clearly labelled as such within the Services.
• Where conversational AI interfaces are provided, they identify themselves as AI-powered in all interactions.

3. Human Oversight

All deployments of our Services operate under a graduated autonomy framework that ensures appropriate human oversight at every stage.

• At baseline, AI provides analysis and recommendations while humans make all decisions.
• Higher levels of AI autonomy are available only through separate written agreement with Nubio, following completion of an operational readiness assessment.
• Safety-critical decisions always require qualified human review and authorisation, regardless of autonomy level.
• Customers select their governance level per workflow, ensuring the right balance of autonomy and oversight for each use case.

4. Data Practices

• Customer data is processed only for the purposes specified in the applicable agreement.
• No customer data is used to train models for other customers without explicit written consent.
• We apply data minimisation principles — we process only what is needed to deliver the contracted service.
• Anonymisation and aggregation are employed where practicable to reduce the use of identifiable data.
• Data processing practices are governed by our Data Processing Agreement.

5. Fairness

• We monitor for algorithmic bias in outputs that may affect pricing, resource allocation, or service delivery.
• We are committed to equitable outcomes — our AI should not produce systematically unfair results based on protected characteristics under applicable law.
• Customers may request information about fairness measures applicable to their deployment.

6. Limitations

• AI outputs are recommendations, not guarantees. They are intended to support, not replace, human judgment.
• Model accuracy depends on the quality of input data. Nubio works with customers to ensure data quality but cannot guarantee the accuracy of source data.
• AI cannot fully account for unprecedented events. Causal modelling mitigates but does not eliminate this limitation.
• Customers retain responsibility for all decisions made using AI outputs.

7. Accountability

• Nubio maintains internal governance processes for overseeing AI development and deployment.
• We review our AI practices against evolving regulatory requirements in Singapore and other applicable jurisdictions.
• We welcome feedback on our AI practices from customers, regulators, and other stakeholders.

8. Contact