AI Transparency Statement

Balian AI advises organisations on governing AI responsibly, and we hold ourselves to the same standard we recommend to clients. This statement explains how we use AI in our own practice and on this Site. It is informed by Singapore’s IMDA Model AI Governance Framework for Agentic AI and, where relevant, the transparency obligations of the EU AI Act (Article 50).

1. Where AI is used on this Site

• The Site itself is a static informational website and does not currently deploy a chatbot, virtual assistant, or other AI system that interacts directly with visitors.
• If, in future, the Site adds an AI-driven interactive feature (e.g. a chat assistant), we will disclose this clearly at or before the first interaction, consistent with Article 50(1) of the EU AI Act and Singapore’s transparency expectations for agentic systems, informing users they are interacting with an AI system, what it is authorised to do, and how to escalate to a human.
• We do not use AI on the Site to generate synthetic media content (images, audio, video, or deepfakes) presented to visitors.

2. Where AI is used in our advisory work

Balian AI designs, and in engagements may deploy or assess, governed AI systems on behalf of clients — including our own governed AI architecture referenced elsewhere on the Site, in which every proposed action is authorised before it executes. In that context, we are guided by the following principles, drawn from Singapore’s Model AI Governance Framework for Agentic AI and our own Responsible AI practices:

• Least agency — AI systems we design analyse and recommend rather than act independently by default; material actions require a named person’s approval.
• Identity and accountability — Where agentic systems are deployed, our designs make each agent’s actions attributable to a defined mandate, a responsible person, and a tamper-evident audit record.
• Transparency to end users — Individuals who interact with an AI system we help deploy, or whose behaviour it affects, are entitled to know: that they are dealing with an AI, what it is authorised to do, how their data is used, and who to contact if something goes wrong.
• Training for internal users — Employees of a client organisation who use or supervise an AI system are given guidance on its failure modes (including hallucination and error-looping) and on maintaining a feedback loop to catch and correct mistakes.
• Human oversight and limits — Systems are designed to pause or escalate near defined risk thresholds rather than relying solely on continuous human vigilance.
• Data boundaries — Personal and confidential data is masked, removed, or restricted before it reaches any component that does not need it. Components that must read source documents are confined to the client’s private environment.

These principles inform our advisory recommendations to clients; they describe how we design and evaluate AI governance, not a guarantee about the behaviour of any specific third-party AI product.

3. Our position on regulatory alignment

We monitor the direction of AI regulation and guidance relevant to our clients, and it informs our advisory practice. It includes:

• Singapore’s Model AI Governance Framework for Agentic AI (IMDA, and related AI Verify testing framework);
• the industry white paper Safeguards for Agentic Finance at Runtime, published with the Monetary Authority of Singapore in July 2026 (a white paper, not regulatory guidance);
• the EU AI Act, including the Article 50 transparency obligations applicable to providers and deployers of certain AI systems (e.g. disclosure of AI interaction, machine-readable marking of synthetic content, and disclosure of deepfakes);
• Korea’s AI Basic Act and Japan’s AI Guidelines for Business; and
• ISO/IEC 42001, the international standard for AI management systems.

4. Human contact

If you have questions about how AI is used in connection with Balian AI or this Site, or wish to raise a concern about an AI-related decision affecting you, contact:
Balian AI (Shasana Balian AI Pte Ltd)
Singapore
Email: info@balian.ai

5. Related policies

See also our Privacy Policy, Terms of Service, and GDPR Statement.

This page is for general information and does not constitute legal advice.