Responsible AI Principles
Softnix is committed to developing and deploying AI responsibly, transparently, and fairly. The framework below describes principles and practices that Softnix applies or is developing across its products; specific controls for each deployment are agreed with the customer.
AI Governance Framework
1. AI Transparency
- Disclose which systems use AI for decision-making
- Explain AI model principles at an understandable level
- Clearly communicate AI limitations
2. Human-in-the-Loop
- Products are designed so that significant decisions and impactful actions can require human review or approval
- Review and override processes are in place
- Final decisions rest with the user
3. Data Privacy in AI
- Softnix seeks to minimize, and where feasible exclude, personally identifiable information in data used to train or fine-tune its models
- Customer data is not sent to third-party AI models without permission
- Data Residency support per organizational requirements
4. Prompt & Data Privacy
- Softnix does not use customer prompts to train models without the customer's authorization
- Session Isolation — each session's data is separated
- Prompt/Response data deleted per Retention Policy
5. AI Risk Management
- Softnix does not currently carry out AI impact assessments, continuous bias/drift monitoring, or model audits as a regular process. In Softnix's own assessment, its AI products do not process personal or sensitive data at a level of risk that meets the criteria for such assessments. This is Softnix's assessment, not a legal opinion, and Softnix may revisit it if its products, their use, or applicable requirements change
- The products are deployed in the customer's environment and the customer decides what data and use cases they are applied to, so customers should assess the risks of their AI use under their own policies and applicable law
- Handling of AI-related incidents through Softnix support channels
6. Responsible AI Principles
| Principle | Description |
|---|---|
| Fairness | AI does not discriminate |
| Accountability | Clear ownership of responsibility |
| Transparency | Explainable and auditable |
| Safety | Safe for users and society |
| Privacy | Respects personal privacy |
Contact
For questions about Softnix AI Governance:
- Email: info@softnix.co.th