ADVOCACY: I-DAIR – The need for trust: artificial intelligence, policy and health
Our digital future is here, and it’s essential that our policies and regulations keep pace with the innovation under way
Digital innovation is inextricably linked to our health and well-being. Health systems are increasingly driven by data. Medical breakthroughs are assisted by machine learning. Patients have greater access through virtual health services. New innovation in artificial intelligence presents immense potential to advance medical treatments, diagnostics and health services alike. We have an opportunity to future-proof health systems, reduce the burden of disease, and shift our focus towards health and well-being through the responsible application of AI and other frontier technologies.
However, AI also poses unique challenges for patients, healthcare workers and policymakers. Questions of ethical data use and privacy are growing, along with public concern over intellectual property, equity and misinformation. The global socio-political environment has contributed to a sense of distrust in both medicine and technology. These challenges are not insurmountable, but they must be addressed to protect individuals while promoting the health of communities around the world.
Policy and regulation matters
We have felt the reverberations of digital technologies in the health sector for decades, but our ability to evaluate, regulate and monitor them has lagged behind. Developments in AI have outpaced existing regulations and policies, necessitating new approaches to the governance of these tools. As both a medical doctor and a former parliamentarian, I have seen firsthand the complexities of developing comprehensive digital and data policies that account for the unique requirements of health systems.
Many of these complexities arise from the competing interests in the digital health space. Policies and regulations must consider the motivations and needs of individuals, healthcare professionals, researchers, public health systems and the private sector. Often, the needs of one group conflict with the motivations of another. Private-sector innovation is an essential force for progress, but may be constrained by the limited resources of many public health systems. Researchers benefit from greater access to data, but individuals must have insight into how their data are managed and shared. These tensions can undermine trust and confidence in artificial intelligence, whereas achieving the right balance can accelerate the positive impacts of AI and other digital health tools.
Policies and regulations must also account for the underlying technological foundation of AI systems. For example, machine learning models can replicate (and exacerbate) biases found in the datasets used to train them. The mainstreaming of cloud computing raises questions around data sovereignty and ownership. Mismatched data standards can create barriers to integration and interoperability. While policy and regulation are not the only remedies to these concerns, they can support an enabling environment that complies with the principles of “First Do No Harm”, one that protects citizens, fosters equity and results in measurable health impacts for the whole population.
Local and global regulation
As with other health regulations, there is a need for both global standards and context-specific policies for the use of artificial intelligence in health. Global standards can help guide governments and ensure protections for the most vulnerable populations. Context-specific policies account for the unique needs of individuals and communities.
For example, the World Health Organization, the International Telecommunication Union and the
World Intellectual Property Organization have created the Global Initiative on AI for Health. The standards set by this initiative will serve as the foundation for country-level regulations and policies that are consistent with global best practices. This initiative will be a cornerstone for country efforts to move from disease-based health systems towards new, more holistic ecosystems for health and well-being.
Every country will require a unique approach to AI policy to align with its specific needs. Countries exist along a continuum of digital transformation and face unique health challenges that cannot be captured in global-level standards. The most material benefits of AI will be felt in low- and middle-income countries that can apply lessons from early adopters to rapidly expand and improve existing health systems. Thoughtful, locally led policy development for AI will catalyse innovation, contribute to greater equity and accelerate health system gains. Upholding country-driven policies will help tailor the use of AI to meet the specific needs of communities everywhere.
Supporting policymakers
In order to create proactive, protective regulation and policy, we must also support policymakers who are responsible for drafting, interrogating and enforcing AI standards. Most of these policymakers are not medical professionals or technologists, so it is imperative that professionals with knowledge and experience in these fields lend their expertise in these critical processes.
Organisations such as UNITE Parliamentarians Network for Global Health bring together policymakers from more than 100 countries to help shape and improve digital health policies. UNITE aims to provide a comprehensive source of advocacy and policy tools that support strong digital health policy and legislation, having already made strides in countries including India, Tanzania and Argentina.
In addition, organisations such as the International Digital Health & AI Research Collaborative can serve as a global regulatory body, supporting countries when designing policies, evaluating tools and preempting regulatory challenges as technology continues to evolve. Working with the WHO and other global initiatives, I-DAIR and its partners can provide a neutral platform to advance AI in health while also ensuring everyone, everywhere is protected and benefits from digital innovation.
Our digital future is happening now, so our policies and regulations must keep pace with digital innovation. We have work to do.
