Why use artificial intelligence?
Like any tool, AI’s potential to be transformative is met by its capacity to be harmful. However, by co-creating a global ecosystem that promotes responsible AI in health care, we can realise this technology’s full potential, equitably and safely
Health care has become a particularly fertile ground for developing artificial intelligence-driven solutions for everything from administrative tasks to population management, clinical decision support tools, virtual assistants and scribes, diagnostics, personalised and precision medicine, surgical and assistive robotics, accelerated drug development, and a growing array of new applications.
A significant share of these technologies has led to measurable improvements in health outcomes. There is even growing evidence of cases where patients have considered ‘machines’ to demonstrate more empathy than their human counterparts. Additionally, AI-driven technologies are becoming increasingly less expensive. Thus, AI has the potential for demonstrable cost savings in health systems where budgets are constantly constrained by growing expenditures and demand.
However, despite its promise, we have yet to see a widespread adoption of AI technologies capable of delivering measurable improvements in health and well-being across societies. This slow improvement is due to the low quality of most health data sets, lack of digital literacy, digital divides, and infrastructural limitations across health systems and communities. But a more fundamental barrier holds humanity back.
Focusing on outcomes
Most AI tools have been implemented by hospitals and clinics with the goal of solving some pressing issues faced by current health systems. And that’s the problem. We continue to see technologies being developed to retrofit AI into current health system architectures. But current models of care are ‘broken’.
The contemporary industrialised healthcare system focuses on processes, not health outcomes. This has led to an increasing burden of disease in communities and ever-growing costs. Current systems limit themselves to reacting to diseases instead of proactively preventing them and promoting health and well-being. This leads to a vicious cycle of the poorest people and those living in the most vulnerable conditions being systematically left behind, and growing as a share of the total population. Continuing to implement AI to feed these disease-reactive systems will inevitably lead to an acceleration of inequalities, a higher burden of disease and unsustainable costs.
So yes, AI has the potential to become the greatest divider or the greatest equaliser of our time. The outcome depends on us.
To embrace AI’s full and positive potential in health care, we need decision makers at different levels of government and management to know exactly what they want to achieve through the application of these technologies. We need leaders to stop seeing AI as a new shiny object, but to ensure that these solutions contribute effectively to a clear vision and purpose.
Why use AI? Will it improve health outcomes and lower the burden of disease? Will it contribute to the sustainability of the health system? These are the questions everyone in power should ask before investing in any AI-driven technology. If the answers are unclear, they should reconsider the proposed investment altogether.
Transitioning to a proactive approach signifies a fundamental shift from hospital-based, disease-centric models to person-centred care. There should be incentives for everyone in the value chain of care to ensure that every action contributes to continuously improved health outcomes that reach all citizens within each community. Such a revolution in healthcare delivery depends on our capacity to perform big data analytics, now made possible by the power of AI. This transformation is essential to lower the burden of disease and foster more sustainable health systems. Such a proactive provision of care prioritises prevention, health promotion and early interventions. By focusing on individual health needs and conditions, care is tailored to the person rather than the disease. Such a shift not only improves health outcomes but also enhances the quality of life for individuals, and promotes equity and accessibility in healthcare services. Moreover, it leads to a more productive and growing economy while fostering a culture of societal and individual well-being. Health as an investment will be perceived by, and benefit, all citizens.
For low- and middle-income countries, AI presents a unique opportunity to leapfrog traditional healthcare development stages and rapidly advance to high-quality, personalised care. These countries can implement cutting-edge AI technologies without the burden and costs of legacy systems, allowing communities in low-resource settings to build efficient tech-driven health infrastructure from the ground up. They can address healthcare disparities, improve access to medical services (including in rural areas) and enhance the overall quality of care, thus contributing to making universal health coverage a reality.
Adopting AI safely
However, AI’s success as a social equaliser will only be possible if the principles of ‘responsible AI’ are followed from the beginning, including in the development of these technologies and their deployment in society. These principles aim to ensure that AI is used in ways that can be considered ethical, equitable, fair, accountable, transparent and sustainable.
To ensure that AI is used responsibly across health systems systematically, we need a governance model with regulatory agencies as gatekeepers, ensuring that these technologies meet rigorous safety, quality and efficacy standards before they gain access to markets.
Post-market surveillance mechanisms are equally necessary to monitor the ongoing performance and impact of AI tools, identifying any unintended consequences and ensuring continuous improvement. A global early warning system can help detect and mitigate potential risks associated with AI deployment, build trust among stakeholders, safeguard public health and, ultimately, accelerate the adoption of these technologies.
For such a governance model to be possible, all stakeholders need to benefit from it, and therefore there must be clear incentives for both the public and private sectors. Health technology assessment and reimbursement models should be updated to reward the value provided by AI-driven solutions. These models must consider the long-term benefits and cost savings associated with the value generated by improved health outcomes and reduced disease burden. Providing financial incentives and support for innovation will stimulate the development and implementation of AI technologies, ensuring that they are accessible and beneficial to all segments of the population.
HealthAI – The Global Agency for Responsible AI in Health is dedicated to accelerating equitable access to AI-driven technologies that improve health outcomes and well-being for all. This mission hinges on building trust among all stakeholders and establishing the right incentives to align efforts to achieve these common goals. Full-scale adoption of AI in health care requires the active participation of governments, healthcare providers, technology developers, academia, patients and citizens-at-large. By joining the HealthAI Community of Practice, you too can play an active role in this transformation. Together, by co-creating a global ecosystem that promotes the responsible use of AI, we can ensure that these technologies contribute to better health for everyone, everywhere. The path forward is clear, and it is time to take decisive action. Let’s get started. ▪
