In a move aimed at shaping the future of healthcare technology, the World Health Organization (WHO) has recently issued a set of guidelines focusing on the ethics and governance of Large Multimodal Models (LMMs).
These models, capable of processing diverse data types like text, images, and video, hold the potential to revolutionize health outcomes.
The guidelines, comprising over 40 recommendations, target governments, technology companies, and healthcare providers, emphasizing the responsible deployment of LMMs to safeguard public health.
Dr. Jeremy Farrar, Chief Scientist at WHO, underscores the importance of transparency in designing, developing, and using LMMs to address health disparities.
Pros and Cons of Large Multimodal Models
WHO’s guidelines highlight five key health applications of LMMs, showcasing their versatility in medicine.
However, they also shed light on potential risks, including the generation of inaccurate or biased information, especially when trained on suboptimal data.
Issues like automation bias, where healthcare professionals may overlook errors, are also flagged.
To ensure the efficacy and safety of LMMs, WHO stresses the involvement of diverse stakeholders, including governments, technology firms, healthcare providers, patients, and civil society, throughout the model’s lifecycle.
Government Roles and Responsibilities
The guidelines propose that governments invest in not-for-profit structures, use regulations to uphold ethical standards, and establish regulatory bodies.
Additionally, they recommend post-publication review and impact assessments to monitor the consequences of deploying LMMs.
Developer Accountability and Best Practices
LMM developers are urged to engage not only scientists and engineers but also potential users and stakeholders in the development process.
The guidelines emphasize the importance of defining tasks with precision, ensuring accuracy, and reliability in LMM performance.
These guidelines build upon WHO’s previous work on the ethics and governance of AI for health, published in June 2021, and offer a comprehensive framework for the responsible integration of LMMs into healthcare practices.
AI’s Impact on Medicine
The guidelines come at a time when AI is making strides in various medical fields, from diagnostics to psychotherapy and drug development.
While showcasing initial success, the WHO guidelines aim to guide the ethical use of AI, ensuring its positive impact on global health.