China Introduces New Ethical Framework for AI-Powered Medical Imaging

China has introduced new ethical guidance for artificial intelligence used in medical imaging, establishing a framework intended to make the rapidly developing technology safer, more transparent and more accountable.

The guidance, prepared by the Medical Ethics Subcommittee of China’s National Science and Technology Ethics Committee and released by the Ministry of Science and Technology, covers AI applications involving medical images such as CT scans, magnetic resonance imaging and ultrasound examinations.

The move comes as hospitals and researchers increasingly explore AI systems that can identify abnormalities in medical images and assist clinicians with diagnosis.

AI Is Becoming More Important in Medical Imaging

Artificial intelligence can process enormous quantities of medical images much faster than a human could examine them individually.

Depending on the system, algorithms can help identify suspicious areas, highlight abnormalities and assist doctors in prioritising cases.

Such technology could potentially improve efficiency and help medical professionals detect certain conditions earlier.

But medical imaging also involves highly sensitive personal information. A patient’s scans can reveal details about their health, making privacy and data security particularly important.

Six Principles Form the Core of the Guidance

The new framework establishes six central ethical principles.

They focus on:

• Promoting human welfare
• Ensuring fairness
• Respecting human autonomy
• Protecting privacy and data security
• Maintaining safety and controllability
• Strengthening transparency

The guidelines place patients’ health and interests at the centre of AI medical-imaging research and application.

These principles are intended to address both technical and human risks associated with increasingly sophisticated medical AI.

Doctors Remain Responsible for Final Decisions

One of the most important provisions is that AI-generated results cannot independently become a patient’s final diagnosis.

The guidance makes clear that human professionals remain ultimately responsible for medical decisions.

AI can provide assistance, identify potential abnormalities or offer analytical information, but doctors retain final authority over clinical decisions.

This approach is designed to ensure that AI remains a supporting technology rather than replacing professional medical judgment.

Patient Consent Becomes a Key Requirement

The new framework also places considerable emphasis on patient autonomy.

Researchers conducting AI medical-imaging studies must obtain informed consent from participants.

Patients must also be allowed to withdraw their consent when they choose, according to the guidance.

This requirement is particularly significant because AI research can involve large datasets containing medical scans and associated patient information.

Giving participants greater control over their involvement is intended to strengthen trust in medical AI research.

Protecting Medical Data

Medical imaging systems can require access to substantial quantities of sensitive information.

CT scans, MRI images and ultrasound results can contain highly personal health details.

The new principles therefore call for stronger protection of privacy and data security throughout the research and development process.

Researchers and developers will need to consider how medical information is collected, stored, processed and used.

Tackling AI Bias

Another major concern addressed by the guidance is algorithmic bias.

AI systems learn patterns from training data. If the data does not adequately represent the populations in which a system will eventually be used, its performance can vary between groups.

For example, an imaging algorithm trained predominantly on one type of patient population could potentially perform less accurately on patients whose characteristics are poorly represented in the training dataset.

The new guidance emphasises algorithmic fairness and calls for efforts to reduce diagnostic bias caused by unbalanced training data.

Developers Cannot Exaggerate Performance

The framework also addresses how AI systems are presented to researchers, healthcare institutions and the public.

It prohibits exaggerating the performance of AI models.

That principle is important because overstating accuracy could encourage medical professionals or patients to place more confidence in an AI system than its evidence justifies.

Transparent reporting of strengths and limitations can help hospitals make more informed decisions about whether and how an AI tool should be deployed.

Safety and Human Control

The guidelines also call for AI medical-imaging systems to remain safe and controllable.

This means researchers need to consider what happens when an AI system produces an unexpected result or performs poorly under conditions that differ from its training environment.

Human professionals should be able to intervene when necessary.

This approach reflects a wider global debate about maintaining meaningful human oversight as AI systems become more capable.

From Research to Clinical Use

The framework is not limited to laboratory research.

It is intended to cover different stages of medical AI development, including algorithm development, trial verification and clinical application.

That makes the guidance relevant to universities, research institutions, hospitals, technology developers and other organisations involved in medical-imaging AI.

The objective is to encourage responsible innovation without allowing rapid technological development to move ahead of ethical safeguards.

Why the Rules Matter

Medical imaging is one of the areas where AI has enormous potential.

A system capable of rapidly analysing thousands of scans could assist doctors in handling growing workloads.

It could also help identify subtle patterns that might otherwise require considerable time to detect.

But an incorrect AI result can have serious consequences when it influences medical decisions.

That is why questions of accuracy, accountability, privacy and human supervision are particularly important in healthcare.

China Is Expanding AI Governance

The medical-imaging guidance forms part of China’s broader effort to establish ethical standards for artificial intelligence.

Chinese authorities have increasingly focused on issues such as data protection, algorithmic fairness, safety, transparency and accountability as AI becomes more deeply integrated into important sectors.

Earlier in 2026, China also introduced broader measures concerning AI ethics review and governance, including requirements covering risk assessment, privacy protection, fairness, transparency and human control.

The medical-imaging guidance applies these concerns to a particularly sensitive area of healthcare.

A Balance Between Innovation and Protection

China’s approach reflects a difficult challenge facing healthcare systems worldwide.

Medical institutions want to take advantage of AI’s ability to process information quickly and potentially improve diagnostic efficiency.

At the same time, patients need assurance that their personal information will be protected and that important medical decisions will not be handed entirely to algorithms.

The new framework attempts to strike a balance between these competing priorities.

The Future of AI in Hospitals

AI-powered imaging is likely to become increasingly common as technology improves.

Future systems may analyse multiple types of medical information simultaneously, assist radiologists with complex cases and help hospitals prioritise patients requiring urgent attention.

As these systems become more capable, the importance of independent testing and continuous monitoring is likely to increase as well.

An AI model that performs well in one hospital or patient population may not necessarily deliver identical results elsewhere.

Human Expertise Remains Central

The new Chinese guidelines make one principle particularly clear: AI is being positioned as a clinical support tool, not an independent medical authority.

Doctors remain responsible for interpreting AI-generated information and deciding what it means for an individual patient.

That approach recognises that medical decisions often require context that cannot be captured completely by an image or algorithm.

A patient’s medical history, symptoms, physical examination and other test results may all influence the final diagnosis.

What Comes Next?

The effectiveness of the new framework will ultimately depend on how consistently its principles are applied in real-world research and healthcare settings.

Hospitals and researchers will need to translate broad ethical requirements into practical procedures for consent, data management, model testing, performance reporting and human oversight.

For AI developers, the guidelines provide a clearer indication of the standards expected when developing medical-imaging technologies.

For patients, the emphasis on privacy, consent and human responsibility provides additional safeguards as AI becomes more involved in healthcare.

China’s new framework therefore represents another step in the evolving relationship between artificial intelligence and medicine.

The technology may transform how doctors analyse medical images, but the new rules make clear that innovation must remain tied to patient safety, ethical responsibility and human decision-making.