The World Health Organization (WHO) has recently released essential regulatory guidelines for the health sector regarding artificial intelligence (AI). This publication emphasizes the need for safety and effectiveness in AI systems, the rapid availability of appropriate systems, and fostering dialogue among various stakeholders such as developers, regulators, manufacturers, health workers, and patients.
With the increasing availability of healthcare data and advancements in analytic techniques, AI tools have the potential to transform the health sector significantly. The WHO recognizes the possibilities of AI in enhancing health outcomes, including strengthening clinical trials, improving medical diagnosis, treatment, self-care, person-centered care, and supplementing the knowledge and skills of healthcare professionals. For instance, AI could be particularly beneficial in settings where there is a lack of medical specialists, such as interpreting retinal scans and radiology images.
However, the rapid deployment of AI technologies, including large language models, without a full understanding of their performance, can have both positive and negative impacts on end-users, including healthcare professionals and patients. AI systems utilizing health data may have access to sensitive personal information, which necessitates robust legal and regulatory frameworks to safeguard privacy, security, and data integrity.
Dr. Tedros Adhanom Ghebreyesus, WHO Director-General, acknowledges the great potential of AI for health but also emphasizes the challenges it poses, including unethical data collection, cybersecurity threats, and the amplification of biases or misinformation. The new WHO guidelines aim to assist countries in effectively regulating AI, harnessing its potential in areas like cancer treatment and tuberculosis detection while minimizing risks.
To manage the rapid rise of AI health technologies responsibly, the publication outlines six key areas for regulation. AI systems are complex and rely not only on the code they are built with but also on the training data they use, which is derived from clinical settings and user interactions. Proper regulation can help mitigate the risks of AI amplifying biases present in training data.
For example, AI models may struggle to accurately represent the diversity of populations, leading to biases, inaccuracies, or failure. Regulations can help address this issue by ensuring that attributes such as gender, race, and ethnicity are reported in the training data, and datasets intentionally include representative samples.
The purpose of the new WHO publication is to provide guidance to governments and regulatory authorities, enabling them to develop new guidelines or adapt existing ones at national or regional levels concerning AI.
In conclusion, the release of WHO’s essential AI regulatory guidelines for the health sector highlights the potential benefits and challenges associated with AI in healthcare. By establishing effective regulations, countries can harness the power of AI while minimizing risks and ensuring the protection of privacy, security, and integrity. This guidance serves as a valuable resource for governments and regulatory authorities as they navigate the complex landscape of AI in the health sector.