Response to Feedback from Public Consultation on the Proposed Exemption from Manufacturer’s Licensing and Product Registration Requirements for Artificial Intelligence

- Software as a Medical Device Developed by Selected Public Healthcare Entities for Use in Public Healthcare

 

 1         The Health Sciences Authority (HSA) invited the public to provide feedback on the proposed exemption from manufacturer’s licensing and product registration requirements for Artificial Intelligence (AI) - Software as a Medical Device (SaMD) developed by selected public healthcare entities for use in public healthcare (“AI-SaMD sandbox”). The public consultation commenced on 19 May 2025 and closed on 19 June 2025.

 2         The responses received were generally positive and supported the proposed AI-SaMD sandbox. Suggestions on the sandbox were also given, as summarised in the table below.

 Table: Summary of suggestions 

Category	Brief description of suggestions received
A	Suggestions to strengthen the controls on AI-SaMDs in the sandbox, for example, by imposing additional International Organisation for Standardisation (ISO), International Electrotechnical Commission (IEC) and cybersecurity requirements, and for the AI-SaMDs in the sandbox to eventually be registered with the HSA.
B	Suggestions to expand the scope of the AI-SaMD sandbox, including allowing other healthcare professionals (apart from registered medical practitioners) to provide oversight over the development of the AI-SaMDs, and extending participation to other parties beyond the public healthcare sector, such as private healthcare institutions and software developers.

 3         The proposed sandbox seeks to strike a balance between ensuring appropriate regulatory oversight and fostering healthcare innovation. In particular, the controls for the sandbox have been developed, taking into consideration the existing cyber and data security safeguards that all public healthcare entities are already required to comply with, and the prevailing regulatory requirements in the Health Products Act and the Healthcare Services Act. Thus, we will:

 a. scope the sandbox to only AI-SaMDs that pose lower potential patient risks, i.e. only Class A and Class B AI-SaMDs, which aim to diagnose and/or drive clinical patient management for only non-critical medical conditions;

b. require that a clinician employed in a public healthcare institution (PHI) and holding the position of Consultant or higher oversees the AI-SaMD's design, validation and output, ensuring that the AI-SaMD is manufactured under the supervision of a qualified medical practitioner;

c. require public healthcare developers to conduct yearly Quality Management System (QMS) self-attestations to confirm that they meet the ISO 13485 standards, similar to the approach for Class A SaMDs under the Health Products Act, to ensure that they maintain a robust QMS;

d. require the AI-SaMDs to be endorsed by the Chairman of Medical Board or Chief Executive Officer of the PHI or Cluster for deployment, ensuring that senior leadership is accountable and aware of the AI-SaMDs deployed in their PHI or Cluster;

e. require that the PHI or cluster notify HSA of the AI-SaMDs before deployment for use in patients;

f. require that deployers must inform patients when AI-SaMDs developed under the proposed sandbox will be used in their medical care; and

g. engage public healthcare developers on HSA product registration upon wider deployment of the AI-SaMDs across public healthcare.

 4         As the sandbox is an expansion of the current exceptions within the Health Products Act, key safeguards, such as post-market obligations and accountability, will continue to be mandated by HSA to ensure equivalent oversight, without compromising patient safety. The requirement to report any adverse effect to HSA, for example, will also continue to apply, and breaches can be subject to enforcement actions, which include fines of up to $20,000 or imprisonment up to 12 months, or both.

5         The AI-SaMD sandbox represents a strategic first step towards broader AI adoption and scaling up across Singapore's healthcare ecosystem. We are starting with the public healthcare sector to gather valuable insights, and will consider extending the scope. The outcomes from the sandbox will be carefully evaluated to determine the next steps, including whether and how the controls should be calibrated, with the aim of creating a comprehensive framework that supports AI innovation while ensuring the safe and effective use of AI in healthcare.

6         HSA thanks all stakeholders who participated in this public consultation on the proposed AI-SaMD sandbox.

 

 HEALTH SCIENCES AUTHORITY
13 FEBRUARY 2026

Published: 13 Feb 2026

Public Consultations