Rapid adoption of artificial intelligence is transforming health research, but WHO warns that existing oversight systems may not fully address emerging risks around privacy, bias, fairness and accountability

The World Health Organization (WHO) has called for stronger and more coordinated ethical oversight of artificial intelligence-related health research, warning that the rapid development and deployment of AI could create risks that existing research-governance systems are not always equipped to manage.

In a report released in September 2026, WHO said AI is creating new opportunities to accelerate scientific discovery, generate health insights and support the development of new technologies. At the same time, the speed and scale of AI adoption are creating new challenges involving privacy, transparency, bias, fairness, accountability and potential harm to individuals and communities.

The report, titled Artificial intelligence-related health research: ethics review and oversight, examines how ethical safeguards can evolve alongside increasingly sophisticated AI applications in healthcare and biomedical research.

AI is changing the pace of health research

AI can process large datasets, identify patterns and assist researchers with tasks that previously required significant amounts of time and human resources. This has the potential to shorten the journey from research and experimentation to real-world healthcare applications.

However, WHO said the faster research-to-product cycle can also create new risks. If weaknesses in an AI system are not identified during the research stage, they may move rapidly into tools used by healthcare professionals, institutions or patients.

The organisation therefore emphasised that ethical assessment needs to keep pace with technological development rather than being treated as a one-time exercise.

Three broad categories of AI health research

WHO's report divides AI-related health research into three broad categories.

The first involves health-related research using AI with data, including the use of artificial intelligence to analyse large datasets and identify patterns associated with diseases or health outcomes.

The second involves research conducted with AI tools and technologies, where AI may support activities such as generating hypotheses, identifying research participants or producing research outputs.

The third covers research on AI tools and technologies, which evaluates the safety, effectiveness, utility and implementation of AI systems used in healthcare.

WHO said these categories can raise different ethical questions, making a single standardised approach to oversight insufficient for every type of AI-related study.

Privacy risks go beyond traditional data protection

The use of health data is one of the central concerns highlighted by WHO.

Even when datasets have been anonymised or information has been made publicly available, AI systems may be capable of identifying previously unexpected relationships or patterns. This means that the original purpose for which data was collected may not fully capture the potential uses and consequences of subsequent AI analysis.

The report therefore points to the need for researchers and institutions to consider how data could be used, combined or interpreted throughout the research lifecycle.

Questions around informed consent, data governance, privacy and appropriate secondary use become particularly important as AI systems become increasingly capable of extracting insights from large and diverse datasets.

Bias and fairness remain major challenges

AI systems learn patterns from data, which means that weaknesses or imbalances in the underlying datasets can influence their outputs.

A system developed using data from one population may not perform equally well when applied to another population or healthcare setting. Differences in demographics, disease patterns, healthcare access and data quality can all affect how an AI model performs.

WHO therefore highlights fairness and equity as important considerations in AI-related health research. Research findings and technologies should be assessed for their relevance across the populations in which they may eventually be used.

Research ethics committees need new capabilities

Research ethics committees (RECs) remain an important part of protecting research participants, but WHO said they may require additional expertise, training and resources to evaluate increasingly complex AI studies.

AI research can involve technical systems whose behaviour may be difficult to explain, large datasets with complicated governance arrangements and algorithms that can change or be updated over time.

These factors can make traditional approaches to assessing research risks more difficult.

WHO's recommendations therefore include strengthening the ability of ethics committees to understand AI-related risks and ensuring that ethical issues are identified at an early stage of research design.

AI research may not always fit traditional review systems

One challenge identified by WHO is that some AI research may not involve direct interaction with human participants.

For example, a project may rely on anonymised or previously collected datasets. In such cases, a study may appear to fall outside conventional human-subject research frameworks.

However, WHO stresses that the absence of direct interaction with individuals does not automatically eliminate ethical concerns. Issues surrounding privacy, discrimination, data ownership, fairness and downstream consequences can remain relevant.

This creates a need for broader governance mechanisms that complement conventional ethics review.

Responsibility must extend beyond ethics committees

WHO's report calls for a wider ecosystem of oversight involving researchers, funders, data-access bodies, scientific journals, professional organisations and regulatory agencies.

The organisation argues that some risks are systemic and cannot be adequately addressed through individual project-level ethics reviews alone.

Researchers are encouraged to identify potential ethical risks early, communicate known limitations transparently and consider the wider social consequences of AI technologies.

Funders and institutions can also influence responsible research by establishing appropriate conditions for data use, transparency and accountability.

Low- and middle-income countries face additional challenges

WHO has placed particular emphasis on the implications of AI research for low- and middle-income countries.

Much of the development of advanced AI technologies and research infrastructure remains concentrated in higher-income economies. This can create imbalances when researchers in other countries contribute data or populations to studies without equivalent levels of local expertise, leadership or participation.

WHO highlights concerns including unequal distribution of benefits, power imbalances, data governance and the possibility of research practices that do not sufficiently involve local institutions and communities.

Building local research capacity and strengthening participation in AI-related health research will therefore be important to ensure that the benefits of the technology are distributed more equitably.

India already has an AI ethics framework for healthcare

India has also established a dedicated framework for addressing ethical issues surrounding AI in biomedical research and healthcare.

The Indian Council of Medical Research (ICMR) published its Ethical Guidelines for Application of Artificial Intelligence in Biomedical Research and Healthcare in 2023. The framework addresses ethical decision-making during the development, deployment and adoption of AI-based healthcare solutions.

The guidelines cover areas including autonomy, safety, privacy, accountability, fairness, informed consent, governance and ethics review. They are intended for stakeholders including researchers, developers, clinicians, ethics committees, institutions, sponsors and funding organisations.

The ICMR has described the guidelines as a living framework that can evolve as AI technology and ethical considerations change.

Transparency and accountability become increasingly important

As AI systems become more deeply integrated into health research, determining who is responsible for decisions and outcomes becomes increasingly important.

Researchers may need to document how AI tools are used, what data they rely upon and what limitations have been identified. Where AI-generated outputs influence research conclusions, appropriate disclosure can also help readers and regulators understand how those conclusions were reached.

Greater transparency can also help identify errors and biases before AI-supported research moves into clinical or public-health applications.

AI can also support ethics oversight

WHO's report does not view AI solely as a source of ethical risk. It also considers how AI itself could potentially assist ethics review and oversight.

AI tools could support researchers and ethics committees by helping organise information, identify potential risks or assist with the assessment of complex research materials. However, such applications would themselves require appropriate safeguards, human oversight and validation.

This creates a dual challenge: ensuring that AI used in healthcare research is governed responsibly while also ensuring that AI used to oversee research does not introduce additional risks.

Public trust becomes a critical factor

The increasing use of AI in healthcare makes public confidence an important component of responsible innovation.

Patients and communities may be more willing to participate in AI-related research when they understand how their data is being used, what protections are in place and who is responsible for the resulting technologies.

WHO therefore links ethical oversight not only to compliance with research rules but also to protecting human dignity, rights, equity and public trust.

Governance will need to evolve with technology

WHO described its latest report as a starting point rather than a final framework for AI-related health research.

The pace at which AI technologies are evolving means that ethical standards and governance systems will also need to adapt. New AI models, data sources and applications may create risks that are difficult to anticipate using existing frameworks.

The organisation's broader position is that innovation and oversight need to develop together so that faster technological progress does not come at the expense of safety, rights and equity.

Market Outlook

The WHO's latest report reinforces the growing importance of responsible AI governance across the healthcare technology ecosystem. As AI adoption expands in medical research, diagnostics, drug development, health-data analytics and clinical applications, companies and institutions may face increasing requirements around data governance, transparency, validation and accountability.

For the healthcare and technology sectors, the development of stronger ethical frameworks could increase the importance of compliance, explainable AI, privacy-preserving technologies, research-quality data and independent validation.

India's existing ICMR framework provides a domestic reference point, while the latest WHO recommendations indicate that AI-related health research governance is likely to remain an evolving global policy area. The key challenge will be maintaining the pace of innovation while ensuring that new AI-enabled health technologies are developed and evaluated with appropriate safeguards for patients, researchers and the wider public.