As recent FT analysis highlights, corporate America may be waxing lyrical about the promise of artificial intelligence but few boardrooms appear able to describe how the technology is actually changing their businesses for the better. There is, however, one sector where the gains are clear, even if it is less eye-catching to profit-chasing investors: public health. For an industry with intensely high demands on accuracy and efficiency, generative AI could transform healthcare delivery and patient outcomes. In turn, the potential benefits for society, the economy and stretched public budgets are immense.
The greatest payback from AI may well come from the earlier and more accurate detection of life-threatening illnesses. In June, Microsoft claimed it had built a diagnostic medical tool that was four times more successful than doctors at determining complex ailments. Some models may even be powerful enough to ascertain distant health risks. Last month, scientists using the gen-AI system Delphi-2M, which was built at the European Molecular Biology Laboratory in Cambridge and trained on large-scale health records, reported that it could predict susceptibility to more than 1,000 diseases decades into the future.
But AI’s impact extends well beyond preventive support. In hospitals, the technology can rapidly analyse X-rays, CAT scans and MRIs. Robotic surgery systems powered by AI can improve surgical precision. Labs are harnessing large language models to accelerate drug discovery too. Crucially, all these applications complement health professionals and free them to provide better care to more patients.
Less glamorous but equally significant is AI’s ability to cut administrative burdens. The US-based Commonwealth Fund estimates that paperwork costs, linked in part to onerous insurance checks, could account for about 30 per cent of America’s excess per capita health spending compared with other nations. In surgeries and hospitals, speech-processing technologies can also be used to transcribe conversations with patients, create structured medical notes and draft letters. A recent study by London’s Great Ormond Street Hospital found a more than 50 per cent reduction in documentation time for clinicians using so-called ambient voice technologies.
Clearly, accelerating AI adoption should be a priority for governments worldwide. Ageing populations and the growing prevalence of chronic diseases in advanced economies are contributing to rising healthcare costs. This strains already stretched public budgets and makes it harder for individuals to afford private cover, as the ongoing wrangling between Republicans and Democrats over insurance support reflects. The World Health Organization also projects a shortage of around 11mn healthcare workers by 2030, which will be more pronounced in lower and middle-income countries. As the Covid-19 pandemic demonstrated, a healthier international population helps reduce the spread of disease, benefiting everyone.
For all the upsides, the use of AI in healthcare is still nascent and patchy. This is partly because the technology is still developing, and rigorous testing is needed before its widespread use in medicine. Health professionals need to train with it too. Data privacy concerns, fragmented sharing networks and outdated IT systems add further complications. Managers can also be reticent over introducing AI where staff may feel their jobs are under threat, or in private systems where revenue depends on high service volumes.
A concerted push from governments, health regulators and tech companies is needed to help fund, trial and deploy AI applications in hospitals, and overcome the cultural and technical implementation obstacles. This will not be easy, but the scale of the prize — in the form of healthier populations and more sustainable healthcare systems — ought to focus the minds.