Hospital AI Lab Practical information on medical AI for Singapore hospitals

Updated:

Generative AI in Singapore hospitals: four risks AIHGle 2.0 names

Hallucination, harmful content, data disclosure and adversarial prompts: what the Ministry of Health’s guidelines say a hospital should do about generative AI tools.

Staff already try chatbots to draft letters, summarise notes and translate. The Ministry of Health’s AI in Healthcare Guidelines 2.0 (March 2026) have a chapter on generative AI that tells hospitals what to expect and what to put in place. For administrators and nurse leaders, it is a ready outline for a staff policy.

Four amplified risks

The guidelines name four risks that generative AI amplifies:

  1. Hallucination. Models may generate and present incomplete or fictitious content as facts, and users may not notice.
  2. Undesirable content. Models may produce harmful content that could compromise clinical outcomes or patient safety, for example a mental health chatbot that responds dismissively to distress.
  3. Data disclosure. Models could leak sensitive data when users submit it, or when the model is prompted in specific ways.
  4. Vulnerability to adversarial prompts. Deliberate attempts to manipulate the model’s output or behaviour can lead to misinformation and security breaches.

What deployers should do

For hospitals, the guidelines say deployers should:

  • design workflows that support fact-checking of outputs;
  • train healthcare professionals on responsible use of generative AI, including “redlines”, the things never to do;
  • support patient communication to address concerns about generative AI;
  • where possible, show clear front-end labels where content is AI-generated;
  • use testing strategies such as red teaming to improve safety and performance.

Users, meaning healthcare professionals, should be trained to evaluate outputs critically and keep clinical judgment as the primary decision-making tool.

Turning it into a staff policy

A one-page policy can follow the same headings. Which tools are approved? Which data types may be typed into each? Who checks drafts before they are used or entered in a record? Where is AI content labelled? Who gets the report if patient data went into an unapproved tool? The templates on this site give a starting draft.

The guidelines also say, more generally, that hospitals should specify the permissible data types for each AI solution, taking into account their data security classification.

Key takeaways

  • Generative AI amplifies hallucination, undesirable content, data disclosure and adversarial prompts.
  • Hospitals should build fact-checking into workflows, train staff, support patient communication and label AI content.
  • Clinical judgment remains the primary decision-making tool.
  • This is general information, not legal advice. The guidelines are guidance, not a statute.

Sources: Ministry of Health and HSA, AIHGle 2.0 (sections 6.1.1(d) and 8.2). Read on 14 November 2026.

This article is part of a series

All series

← Blog