Artificial intelligence (AI) holds much promise for healthcare, but comes with the challenges of amplifying biases or misinformation, cybersecurity, and so on, says a World Health Organization publication outlining the regulatory considerations related to AI for health. 

With the increasing availability of healthcare data and analytical techniques — whether machine learning, logic-based or statistical — AI tools could transform the health sector, the UN health agency said.

However, “AI technologies — including large language models — are being rapidly deployed, sometimes without a full understanding of how they may perform, which could either benefit or harm end-users, including healthcare professionals and patients. When using health data, AI systems could have access to sensitive personal information, necessitating robust legal and regulatory frameworks for safeguarding privacy, security, and integrity, which this publication aims to help set up and maintain,” it added.

Pointing to challenges including unethical data collection, cybersecurity threats, and amplifying biases or misinformation, WHO Director-General Dr Tedros Adhanom Ghebreyesus said that the new guidance would “support countries to regulate AI effectively, to harness its potential, whether in treating cancer or detecting tuberculosis, while minimising the risks.” 

Managing AI responsibly 

On AI’s potential to improve health outcomes, the note pointed to strengthening of clinical trials; improving medical diagnosis; and supplementing healthcare professionals’ knowledge and skills. In fact, in places that lack medical specialists, AI can help in interpreting retinal scans and radiology images, among many other uses, it said.

The publication outlined six measures on managing AI health technologies responsibly.

To foster trust, the publication stressed on transparency and documentation, such as documenting the entire product lifecycle and tracking development processes. For risk management, issues like ‘intended use’, ‘continuous learning’, human interventions, training models, and cybersecurity threats must all be comprehensively addressed, with models made as simple as possible, it said.

Externally validating data and being clear about the intended use of AI helps assure safety and facilitate regulation. A commitment to data quality, such as through rigorously evaluating systems pre-release, is vital to ensuring systems do not amplify biases and errors. 

The challenges posed by important, complex regulations — such as the General Data Protection Regulation (GDPR) in Europe and the Health Insurance Portability and Accountability Act (HIPAA) in the US — are addressed with an emphasis on understanding the scope of jurisdiction and consent requirements, in service of privacy and data protection, the note said. Fostering collaboration between regulatory bodies, patients, healthcare professionals, industry representatives, and government partners can help ensure products and services stay compliant with regulation throughout their lifecycles, it added.