Human-Machine Partnership

In Deira, just steps away from the headquarters of Dubai Future Foundation and its creative spaces, and near the center of decision-making, the ninth edition of the “Zaman” Digital Health Conference kicked off. We had the honor of moderating discussions with a select group of international experts on the future of electronic health records, the upskilling of healthcare personnel, and the competencies required to keep pace with digital transformations. It was hardly surprising that artificial intelligence featured in nearly every dialogue and presentation, given the opportunities it offers and the preparedness it demands.

Among the proposals that caught our attention was a call by Mr. Brian DeFrancesca to create a new role within the healthcare sector. The incumbent would break down daily work into minute tasks, understand the requirements of each task, and determine which tasks can be delegated to artificial intelligence. Taking nursing as an example, a nurse’s duties include collecting, documenting, reviewing, and coordinating information with other teams. However, when technology enables the nurse to perform some of these tasks, they may find more time to listen to a hesitant patient revealing their concerns, or to notice changes in a patient’s condition that warrant medical intervention.

If we deconstruct the work of a researcher in the same way we deconstruct that of a nurse, we can see where artificial intelligence assumes specific tasks and where the nature of the task requires that decision-making remain in human hands. This is illustrated by a recent study published in Nature, which describes a laboratory developed by researchers from Shanghai Jiao Tong University in China to plan and execute specific experiments involving DNA. An experiment begins with a textual prompt in which the researcher specifies their objectives. The system then coordinates the work of machines and robots until the experiment is completed, while the human team intervenes to maintain equipment and replenish necessary materials.

Imagine what this means for a researcher who spends hours preparing samples and repeating procedures: they may now have room to pursue an unexpected result that might otherwise have remained a mere note in a lab notebook, potentially leading to a question worthy of a new experiment. However, delegating execution does not absolve us of the need to understand science and the fundamentals of experimentation; rather, it heightens the importance of deep training in these areas. If we merely request experiments and receive their results, we may gradually lose the expertise that enables us to distinguish logical outcomes from errors, and to detect flaws in experimental design or interpretation. Algorithms may propose incorrect steps or offer convincing explanations that conceal errors or hallucinations, amplifying the danger when we are unable to critique them.

Therefore, we need to prepare researchers who understand how results were generated and can verify them, with clear accountability for researchers and institutions to review and endorse findings, regardless of the autonomy granted to machines. We also need to review our healthcare, research, and service operations, identify tasks suitable for delegation, and test them incrementally. This should be accompanied by training that enhances evaluation, interpretation, and communication skills, along with safeguards for data protection and result verification. We must measure impact in terms of time available for patient care, research presence, and improved beneficiary experience, while preserving the expertise that allows us to steer and hold these systems accountable.

Undoubtedly, human-machine partnership opens broad horizons for science and practice, offering us the opportunity to direct our time and capabilities toward what requires deeper thinking and greater attention. The better we choose what to delegate to machines, and the more we retain knowledge, guidance, and responsibility, the more we can build a partnership that advances science and grants humans a wider space for creativity.

Dr. Dhafer Adel Al-Huwail