Artificial Intelligence May Be ‘Threat’ to Human Health, Experts Warn

In a recent analysis published in BMJ Global Health, an international group of researchers and public health experts have argued that artificial intelligence (AI) and artificial general intelligence (AGI) may pose numerous threats to human health and well-being, calling for research into these technologies to be halted until they can be properly regulated.

The authors noted that AI technology has various promising applications in healthcare, but posit that misuse of these solutions could harm human health through their impact on social, economic, political, and security-related determinants of health.

The research and development of healthcare AI are progressing rapidly, the authors stated, highlighting that much of the literature examining these tools is focused on the potential benefits gained through their implementation and use. Conversely, discussions about the potential harms of these technologies are often limited to looking at the misapplication of AI in the clinical setting.

However, AI could negatively impact ‘upstream’ determinants of health, characterized by the American Medical Association (AMA) as individual factors that may seem unrelated to health on the surface, but actually have ‘downstream’ impacts on patients’ long-term health outcomes.

The AMA indicates that these upstream factors, such as living conditions or social and institutional inequities, have not always been within the scope of public health research but can exacerbate disease incidence, injury rates, and mortality.

The authors argued that the potential misuse and ongoing failure to anticipate, adapt to, and regulate AI’s impacts on society could negatively affect these factors and cause harm.

The analysis identified three impacts AI could have on upstream and social determinants of health (SDOH) that could result in threats to human health: the manipulation and control of people, the proliferation of lethal autonomous weapons systems (LAWS), and the potential obsolescence of human labor.

The first threat, the authors explained, results from AI’s ability to process and analyze large datasets containing sensitive or personal information, including images. This ability could enable the misuse of AI solutions in order to develop highly personalized, targeted marketing campaigns or significantly expand surveillance systems.

These could be used with good intentions, the authors noted, such as countering terrorism, but could also be used to manipulate individual behavior, citing cases of AI-driven subversion of elections across the globe and AI-driven surveillance systems that perpetuate inequities by using facial recognition and big data to produce assessments of individual behavior and trustworthiness.

The second threat is related to the development and use of LAWS, which can locate, select, and engage human targets without supervision. The authors pointed out that these can be attached to small devices like drones and easily mass-produced, providing bad actors with the ability to kill “at an industrial scale.”

The third threat is concerned with how AI may make human jobs and labor obsolete. The authors acknowledged that AI has the potential to help perform jobs that are repetitive, unpleasant, or dangerous, which comes with some benefits to humans. However, they noted that currently, increased automation has largely served to contribute to inequitable wealth distribution and could exacerbate the adverse health effects associated with unemployment.

In addition, the authors described how AGI could pose an “existential threat to humanity.”

“We are now seeking to create machines that are vastly more intelligent and powerful than ourselves,” they said. “The potential for such machines to apply this intelligence and power—whether deliberately or not—in ways that could harm or subjugate humans—is real and has to be considered.”

They highlighted that AGI’s connection to the internet and the real world, including robots, vehicles, digital systems that help run various aspects of society, and weapons, could be “the biggest event in human history,” for the benefit of humanity or to its detriment.

Because of the scale of these potential threats and the significant impacts they could have on human health, the authors stated that healthcare professionals have a critical role to play in raising awareness around the risks of AI. Further, the authors argued for the prohibition of certain types of AI and joined calls for a moratorium on AGI development.

“With exponential growth in AI research and development, the window of opportunity to avoid serious and potentially existential harms is closing. The future outcomes of the development of AI and AGI will depend on policy decisions taken now and on the effectiveness of regulatory institutions that we design to minimise risk and harm and maximise benefit,” they wrote.