The World Health Organization (WHO) has issued a stern warning regarding the rapid deployment of Artificial Intelligence (AI) technologies, especially large language models, without a comprehensive understanding of their potential impacts on end-users, including healthcare professionals and patients.
This caution came as WHO unveiled a set of guidelines for the global regulation of AI in healthcare. The organization emphasized the complexity of AI systems, highlighting that their performance depends not only on the underlying code but also on the data used for their training.
Such data originates from clinical settings and user interactions, which may not always accurately represent the diversity of populations, potentially leading to biases, inaccuracies, or even system failures.
WHO stressed that the utilization of health data by AI systems can entail access to highly sensitive personal information, necessitating robust legal and regulatory frameworks to protect privacy, security, and data integrity.
Dr. Tedros Adhanom Ghebreyesus, the WHO Director-General, expressed the promise and challenges of artificial intelligence in healthcare, including concerns related to unethical data collection, cybersecurity threats, and the amplification of biases and misinformation.
The newly released guidance aims to assist countries in effectively regulating AI, allowing them to harness its potential in healthcare applications such as cancer treatment and tuberculosis detection while minimizing associated risks.
The publication outlines six key areas for AI regulation in healthcare, including transparency and documentation, external data validation, and the clear definition of AI’s intended use.
Furthermore, the WHO publication underscores the importance of fostering collaboration among regulatory bodies, patients, healthcare professionals, industry representatives, and government partners. Such collaboration can help ensure that AI products and services remain compliant with regulations throughout their life cycles.
Despite the warnings, the statement acknowledges that well-regulated AI tools hold the potential to enhance health outcomes significantly. This can be achieved by strengthening clinical trials, improving medical diagnoses and treatments, promoting self-care, and providing person-centered care.
AI can also supplement healthcare professionals’ knowledge, skills, and competencies, particularly in settings with a shortage of medical specialists, such as in the interpretation of retinal scans and radiology images.