Put patients at the centre of medical AI governance

In a recent article published in Nature, Kyle Lam and his co-authors have proposed a framework for categorizing medical artificial intelligence (AI) based on autonomy, automation, and operational scope. These three properties were found to be essential in establishing regulatory boundaries and liability mechanisms across various capability levels of medical AI.

The framework emphasizes the importance of allocating responsibility to clinicians, healthcare organizations, device manufacturers, and regulators. However, the authors acknowledge that patient rights have not been systematically integrated into the design of this framework. This omission is critical, as the success of medical AI largely depends on patient adherence and acceptance.

One of the key aspects of the framework is its potential application to AI models that predict the effectiveness of breast-cancer drugs. By accurately predicting which drugs will work best, these AI models can significantly improve patient outcomes and reduce the trial-and-error approach in treatment selection. This can potentially save time, reduce costs, and enhance the overall quality of care for patients.

Another interesting aspect discussed in the article is the concept of the "Einstein test," which explores what would happen if AI were to attempt to rediscover fundamental theories like relativity. This thought experiment highlights the potential for AI to revolutionize not only medical treatments but also our understanding of the universe.

The article also briefly mentions the risks associated with clinical-trial outcomes being placed on prediction markets. While such markets could potentially increase transparency and efficiency, they also introduce new risks and uncertainties that need to be carefully managed.

Finally, the article pays tribute to recent tragic incidents where two children died from gene therapies in China. This serves as a sobering reminder of the need for ongoing vigilance and strict regulatory oversight in the rapidly evolving field of biomedical technologies, ensuring patient safety remains the top priority.