An Overview of Recent Digital/Artificial Intelligence (AI) Guidelines in the Asia Pacific Region

Introduction

AI is everywhere these days—it’s impacting everything from how we plan our vacations, to what we see online, and perhaps how we attack our email inbox. How is it impacting drug development, and how are global regulatory agencies adapting to the use of AI? ISPE’s Regulatory Quality Harmonization Committee’s Asia Pacific Regional Focus Group has reviewed a few recent AI-related guidances from the Asia Pacific region, and is pleased to share the following insights.

Australian Government, Department of Industry Science and Resources, on AI

The Australian government lists AI as a critical technology in the national interest. A plan has been developed over a number of years with AI ethics principles being published as early as 2019. The National AI Plan sets out the government’s direction on AI. It focuses on seizing the potential of AI through:

• Capturing the opportunity
• Spreading the benefits
• Keeping Australians safe

Capturing the AI Opportunity

The aim is to capture the opportunity for Australia to become a global leader in developing and adopting trusted, secure, and responsible AI through programs such as:

• National Artificial Intelligence Centre (NAIC) the government lead body supporting industry
• AI Accelerator Initiative to fund cooperative research centre programs
• Setting expectations for Data Centres to align with national interest
• Collaborating with leading AI and technology companies

Spreading the Benefits

Every Australian should be able to benefit from AI, regardless of age, location, or gender. Spreading the benefits of AI comes from supporting adoption. AI adoption has the potential to improve business productivity and deliver better wages, job satisfaction, and stability for workers. This is done through:

• AI Adopt Program to fund centers to help businesses adopt AI
• Guidance for AI Adoption, including a Voluntary AI Safety Standard and AI Ethics Principles

Keeping Australians Safe

To protect Australians and maintain trust and confidence in AI systems, the government is committed to understanding and responding to advanced AI. Controls include:

• Artificial Intelligence Safety Institute
• AI safety science, through programs which advance local and international AI safety research

Partnering Internationally on Responsible AI

Australia engages internationally on AI to support Australian innovation and to shape international AI governance. Partnerships include:

• Paris AI Action Summit
• Bletchley Declaration on AI Safety
• Seoul Declaration
• Hiroshima AI Process Friends Group
• Global Partnership on Artificial Intelligence (GPAI)
• EU-Australia Digital Economy and Technology Policy Dialogue

Within the Health Sector, as part of this national framework, Australia’s Therapeutic Goods Administration (TGA) has been working to ensure that AI is safe to use in medical devices. Guidance for industry is provided through the TGA web site. Below is a short summary of some of this guidance.

Artificial Intelligence (AI) and Medical Device Software Regulation

Australia’s medical device regulatory framework is technology agnostic which means products are regulated based on their intended purpose, not the technology they use. The intended purpose—what the product is used for—is defined by the manufacturer and will determine whether the product meets the definition of a medical device under the Act regardless of the platform used—whether a watch, phone, tablet, cloud service, laptop, or hardware device.

Software or AI products including apps, websites, programs, internet-based services or packages will be regulated as medical devices if they are intended for:

• Diagnosis, prevention, monitoring, prediction, prognosis, or treatment of a disease, injury, or disability
• Alleviation of, or compensation for, an injury or disability
• Investigation of the anatomy or of a physiological process
• Control or support of conception

Manufacturers are required to monitor how system updates affect their products’ functionality. Where new features or functionality change the intended purpose or performance of a medical device, these must not be implemented until the device has received appropriate regulatory approvals. It is therefore important for the manufacturer to continuously monitor their software’s performance and functionality to assess the impact of updates or identify unintended scope creep.

In addition to general software requirements, the manufacturer of software that uses AI or machine learning (ML) is required to possess evidence that is sufficiently transparent to enable evaluation of the safety and performance of the product.

Synthetic data may be used in place of data from real patients or devices for training or validation of AI systems. Manufacturers must provide a clear rationale for its use, along with a description of how the data was generated and its relevance to the intended use. However, for many use cases, synthetic data may not provide sufficient depth and variability to adequately validate a product. Where data is readily available in large volumes, synthetic data is less likely to be considered appropriate.

The Association of Southeast Asian Nations (ASEAN) Guide on AI Governance

In addition to Australia, ASEAN published a general guide on AI Governance and Ethics in 2024. The goal of the ASEAN association is to promote economic, political, and security cooperation among its members (Brunei, Cambodia, Indonesia, Laos, Malaysia, Philippines, Singapore, Thailand and Vietnam, plus Timor-Leste as a member in principle). As such, this AI guide was collaboratively developed by all ASEAN member states to serve as a best practice for designing, developing and deploying AI technologies in the AP region, and is intended to be a living document that is periodically reviewed and updated. It outlines seven guiding principles:

• Transparency and explainability
• Fairness and equity
• Security and safety
• Human-centricity
• Privacy and data governance
• Accountability and integrity
• Robustness and reliability

These principles ensure that the development of AI in the region takes into account societal impact of the use of the tools. In addition to guiding principles, the guide also describes four key components organizations should implement as they develop a governance framework for their own AI use: internal governance structures and measures, determining the level of human involvement in AI-augmented decision-making, operations management, and stakeholder interaction and communication. Additional resources within the guide include recommendations on implementing AI at the national and regional levels, an AI risk impact assessment template (Annex A) that lists questions to enable developers to systematically consider risks of their AI tools, and six use cases of AI implementation in the APAC region that illustrate how principles in the guide have been implemented.

The guide demonstrates support for the responsible implementation of AI, and supports the overall ASEAN purpose to promote active collaboration and mutual assistance on matters of common interest in the economic, social, technical, and scientific fields.

Singapore on AI Governance

To support responsible adoption of AI, the Singapore Infocomm Media Development Authority issued cross-cutting governance frameworks. These include the Model AI Governance Framework for Agentic AI, which provides “guidance to organizations on how to deploy agents responsibly, while emphasizing that humans are ultimately accountable.” The framework emphasizes risk assessment, human accountability, controls, and transparency. The instruments provided in the framework, while not sector-specific, could shape expectations for how pharmaceutical companies govern AI tools.

Within the health sector, the Ministry of Health (MOH) and the Health Sciences Authority (HSA) jointly issued Artificial Intelligence in Healthcare Guidelines in March 2026. While these guidelines do not regulate pharmaceutical manufacturing specifically, their lifecycle, accountability, and risk-based driven principles provide a regulator-endorsed framework that pharmaceutical manufacturers can adopt as good practice when deploying AI.

Furthermore, HSA has articulated a detailed lifecycle regulatory framework for software and AI/ML-enabled systems through guidance such as Regulatory Guidelines for Software Medical Devices, including Machine Learning-Enabled Medical Devices – A Life Cycle Approach, which—although formally applicable to software medical devices—provides further insights on regulators’ expectations in the manufacturing of pharmaceutical products:

• A central theme of these guidelines for software medical devices is the total product lifecycle approach. Software must be managed from requirements definition through design, validation, deployment, maintenance, change, and eventual retirement, especially when it could impact the quality of the product, such as critical process parameters (CPPs), critical quality attributes (CQAs), and batch disposition decisions.
• For AI systems, based on these guidelines, inspectors will increasingly expect manufacturers to explain why the model’s outputs can be trusted, not merely that the system operates as designed.
• These guidelines explicitly extend risk management beyond traditional software failures to include AI-specific risks with respect to manufacturing, such as system degradation, false negatives allowing quality issues to pass undetected, false positives causing unnecessary batch rejections, and the impact of retraining or data changes on established process knowledge.
• These guidelines emphasize continuous monitoring to detect new risks or degradation after deployment. For self-learning or adaptive models, this element becomes especially critical, as performance changes may be gradual and not immediately visible.

These guidelines demonstrate the support the Singapore government, MOH and HSA offer to the pharmaceutical sector in terms of AI, provided that existing regulatory standards for quality and safety are met (reference article from the Straits Times dated March 2026).

South Korea Ministry of Food and Drug Safety (MFDS) Guidance on AI

Artificial intelligence is rapidly becoming a cornerstone of regulatory science, and South Korea’s MFDS is now embedding AI across its operations. In its 2026 Work Plan, MFDS outlined how AI will reshape safety oversight, regulatory review, and compliance monitoring. This marks a shift from pilot projects to systemic integration, positioning Korea as a global benchmark in AI-enabled regulation.

One of the most visible areas is food and drug safety. MFDS announced the use of AI risk prediction models for imported food inspections, AI detection of foreign substances in meat, and predictive analytics to identify causes of foodborne illness. These initiatives are complemented by the expansion of smart HACCP systems, embedding AI into manufacturing and distribution oversight. This demonstrates a move toward proactive, data-driven safety management.

AI is also being deployed in advertising and online monitoring. The agency introduced “AI캅스”, a real-time monitoring system designed to detect and block illegal online drug sales and misleading health advertisements. MFDS has emphasized that AI will be used to identify fake doctor or pharmacist endorsements, strengthening consumer protection and raising compliance expectations for industry players.

Drug safety and misuse prevention are another priority. MFDS is building an AI-driven integrated monitoring system for narcotics misuse, expanding prescription oversight, and applying predictive analytics to identify abuse trends. The 2026 plan also broadens the scope of self-prescription bans and requires more comprehensive patient prescription history checks. These measures reflect a shift toward preventive regulation, using AI to anticipate risks before they escalate.

Perhaps most relevant for pharma peers is MFDS’s commitment to shorten approval timelines. The agency has set a target to reduce drug approval review periods from 420 days to 240 days by deploying an AI approval review support system. This builds on MFDS’s earlier Guidance on Clinical Trials Design of AI-based Medical Devices (2023) and Guidance on the Review and Approval of AI-based Medical Devices (2023), which established clear pathways for machine learning-enabled devices. Together, these documents provide the technical foundation for AI-assisted regulatory review, ensuring that submissions are transparent, validated, and lifecycle-managed.

These initiatives build on Korea’s Digital Medical Products Act, enacted in January 2025, which established a comprehensive framework for AI-based diagnostics, digital therapeutics, and software-driven medical devices. Combined with Korea’s leadership in hosting international AI regulatory symposia (AIRIS 2024 with the US Food and Drug Administration (US FDA), and AIRIS 2025 with the World Health Organization [WHO]), the country is positioning itself as a global leader in harmonizing AI regulation. For pharma, the implications are clear: dossiers must be AI-ready, compliance must anticipate machine-driven oversight, and engagement with regulators will increasingly involve AI systems.

China on AI

On April 2, 2026, China’s National Medical Products Administration (NMPA) issued the Implementation Opinions on Artificial Intelligence (AI) and Drug Regulation (hereinafter referred to as the “Opinions”). The Opinions outline the plan for an integrated system for drug regulation and artificial intelligence by 2030 and is supplemented by a policy interpretation document which provide the background on the issuance of the Opinions and the key focus of the regulatory reform.

The Opinions propose seven key focus areas to develop the next stage of regulatory digital intelligence. The first key focus is to build a human-machine collaborative intelligent system to improve the efficiency and quality of the review and approval process for drugs, medical devices, and cosmetics. The second focus area is to enhance intelligent supervision capabilities across the entire supply chain from initial research and development and clinical trials to manufacturing, logistics, and post-market distribution.

The AI-enhanced supervision capabilities also extend to a risk monitoring system where advanced data-driven models will be used to monitor complaints and reports, online sales, and high-risk products, to identify risk signals and provide early warnings. This also includes the development of a smart drug testing system involving robotics technology.

The other key focus is to establish a smart, risk-based inspection management platform where inspection plans can be developed based on big data of audit history and records of pharmaceutical and medical device companies. The use of AI to determine inspection targets and frequencies can strengthen the enforcement of pharmaceutical regulations while reducing redundant inspections.

Besides the application of AI in the core regulatory work, the Opinions also discussed the use of digital technologies to enhance collaboration and information-sharing across departments and regions with the development of a national integrated business system. This initiative will also be extended to government services by integrating AI into online consultations and other intelligent assistance services.

By setting clear directions for regulatory reform, NMPA hopes to guide and encourage the industry to accelerate its digital transformation throughout the value chain, especially in the production and testing of high-risk products such as blood products and traditional Chinese medicine.

Besides the key focus areas summarized above, the Opinions also discussed NMPA’s implementation plan to support the rollout of artificial intelligence in drug regulation. This includes not just the construction of high-quality datasets and large-scale model platform but also enhancing the computing infrastructure and security protection system.

In summary, the AI-driven regulatory reform led by NMPA is both broad and targeted, covering most, if not all, aspects of regulatory operations. The vision of NMPA is to create a collaborative ecosystem for smart drug safety governance by 2035, built on digital intelligence and autonomous control.

Conclusion

This blog post summarises the current regulatory expectations from several Asia Pacific (AP) markets regarding the use of artificial intelligence in pharmaceutical development/registration/regulation. Some have approached the situation from a principles perspective (e.g. ASEAN), some have progressed for use in specific cases (e.g. South Korea).

In contrast to the potential divergence amongst AP regulators, regulators in the US and the European Union (EU) have moved towards alignment, most notably through the recent publication of a joint set of ten principles for the safe, ethical, and reliable use of AI in medicines. While these principles remain high-level, they signal a clear intent to harmonise expectations across regions and to transition from conceptual guidance toward practical regulatory application, such as risk-based inspection targeting and AI-enabled review processes.

Looking ahead, there are strong indications that regulatory approaches within the Asia Pacific region will continue to converge with those of the US and EU. Initiatives such as structured content authoring, US FDA’s Knowledge-aided Assessment and Structured Application (KASA), and the transition under ICH M4Q(R2) from document-based to information-based submissions will increasingly necessitate the effective use (and thus supervision) of digital tools, including AI.

As global pharmaceutical development and supply chains rely on common data standards and interoperable regulatory frameworks, Asia Pacific regulators are expected to closely follow and, in some cases, potentially move at pace with these evolutions, facilitating the responsible and transparent use of AI within established quality and safety paradigms.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Disclaimer

iSpeak blog posts provide an opportunity for the dissemination of ideas and opinions on topics impacting the pharmaceutical industry. Ideas and opinions expressed in iSpeak blog posts are those of the author(s) and publication thereof does not imply endorsement by ISPE.