As artificial intelligence (AI) tools make their way into healthcare, researchers argue that for AI to truly transform public health, it must address the underlying social and political determinants of health, not just focus on disease treatment. Public health experts warn that without proper regulation and a shift in priorities, AI could exacerbate existing inequities and fail to create meaningful change.
Systematic Action to Address Health Disparities
A team of researchers at UC Santa Cruz and the University of British Columbia (UBC) has conducted an extensive assessment of the current trajectory of AI in healthcare. They suggest that AI too can transform public health, but could be doomed to repeat the “technological solutionism” of recent decades. A focus on treatment over social, economic, and political determinants of health.
One example is the AI-powered tools under development to manage supply chains, disease outbreaks, or even medical imaging. While apps can help facilitate care and ensure it is timely and streamlined, they do little to impact the actual determinants of health such as poverty, food insecurity, unsafe housing practices, and systemic racism. Rebooting Flint as one of the study’s authors, Lucia Vitale, said: “We continue to invest heavily in these tech silver bullets that never seem to shift public health for real because they’re avoiding dealing with the political and social determinants of health.”
Concerns about Equity and Non-Targeted Effects of Reforms
The researchers also underscore various equity considerations of AI implementation in healthcare. Well, we know in other industries there have been – and will almost certainly continue to be – concerns that AI could reproduce or even worsen existing biases and discrimination. However, AI tools can simply perpetuate inequalities if the data used to train their models is biased (likely due either to historical biases being reflected in the dataset or as a result of the system to be modeled suffering said bias) and could at worst make life harder for different groups like racial minorities, women, etc.
Also… people with preexisting conditions or bad health: fear that an AI tool may be misused by insurance companies or policymakers to discriminate against them. This provides little reassurance as AI could be used to justify denying coverage or after allocating resources in ways that perpetuate the most marginalized communities.
The researchers also caution that the current ownership and profit models of AI may lend themselves to further concentrating wealth in high-income countries and corporations, with data extraction or testing of new, potentially dangerous technologies done in low- or middle-income nations. This is the challenge that any AI-powered healthcare solutions need to address — otherwise, we are generalizing a pattern of exploitation, affecting equitable access and distribution.
Charting a Course: Responsible Governance and Comprehensive Interventions
If the full power of AI can be realized in public health, the researchers conclude, it is vital to begin establishing strong regulatory frameworks and ethical governance structures now before these new technologies are allowed to grow any further. This includes dialogue with civil society, policymakers, and global health organizations such as the World Health Organization to establish clear guidelines and accountability mechanisms.
The economists also argue that AI interventions should augment not replace evidence-based care available through comprehensive services programs like community health worker programs and harm reduction initiatives. Instead, AI could be used to augment general practitioners or provide basic care in under-served regions by strengthening the healthcare system.
The long-term success of AI in public health hinges on the ability of these stakeholders to reframe the priorities and incentives underpinning technological development, redirecting these toward the collective benefit and meeting the needs of those who are most at risk. Co-author of the study Leah Shipton said, “This is our chance to put right how we govern new technologies but we need a plan that sets out what ethical governance in AI health technologies should require.”
