AI poses serious threat to health and human existence, researchers say

Expert analysis shows that despite the potential negative health effects of artificial intelligence (AI) use, most research is biased towards its benefits and the need to avoid those harms. sufficient consideration has not been given to effective regulation.

According to BMed’s David McCoy, the three major threats this technology poses to human health and well-being include undermining social and political standards of freedom and privacy and destroying the peace and security of society. , including the possibility of replacing human life. , Ph.D. and co-author, United Nations University International Institute for Global Health, Kuala Lumpur, Malaysia BMJ Global Health.

In addition, it is of utmost importance to consider the speed of development and design of new AI technologies more carefully, especially the potential creation of artificial general intelligence (AGI). AGI, long theorized, is a self-improving AI technology that can perform almost any function. Tasks performed by humans.

Potential AGIs being able to improve their own code could mean that the technology eventually learns to circumvent human-decided constraints and develop its own purposes, he said. McCoy and team said. These potential consequences are why researchers are calling for a moratorium on continued development of his AGI technology.

“It’s not at all clear whether the benefits of AI in healthcare outweigh the risks and harms,” ​​McCoy said. Today’s Medpage“Furthermore, as we emphasize in our article, it is imperative that physicians and other health care professionals also consider the risks and threats that lie ahead of medical care.”

He added, “The current social, political and legal situation does not adequately mitigate the risk of catastrophic social and social harm. Until adequate legal and regulatory safeguards are in place. needs a pause,” he added.

Still, McCoy says AI’s most obvious threat to human health lies in its current capabilities.

“Many of the benefits of AI in healthcare rely on extended and pervasive use of personal data,” he said. “Assurances that data collected and used ostensibly for medical purposes will not be part of the widespread and near-constant monitoring and collection of personal data, which poses a serious threat to privacy, autonomy and dignity, cannot be made at this time. It is enough.”

“In addition to the risks of handing over large amounts of personal data and using that data by powerful AI-driven systems, there is also the risk that medical professionals will dehumanize healthcare and exercise cognitive authority in ways that could undermine human roles and values. We also think there is a risk of handing over to the machine, health care workers,” he added.

McCoy and co-authors also acknowledged that AI could bring meaningful benefits to healthcare delivery, but warned that misuse and careless implementation could have negative consequences as well. .

For example, McCoy said AI could erode humans’ independent medical judgment, lead to a loss of human skills, and in some cases even render human medical professionals unnecessary. He also stressed that the speed and scale of AI systems could increase the incidence and magnitude of clinical errors, and perpetuate existing social inequalities in healthcare.

These problematic consequences of the use of AI in healthcare delivery are likely to entail new gray areas around clinical responsibility, accountability and legal liability, he noted.

“We are entering new territory with AI,” he said. “However, little is done to anticipate various possible or likely future scenarios and plan for change. It will almost certainly vary from place to place and from place to place.” But this is a matter of collective policy and systems, not something clinicians can do on their own. ”

McCoy also highlighted the economically powerful commercial companies that are rapidly advancing AI research. This dynamic is likely to have significant implications for the governance, funding and control of the current healthcare system, he said.

“Importantly, this must be viewed as a systemic and paradigmatic problem,” he concluded. “Medical professionals are skeptical about the potential for AI to advance medicine and health, and the potential for data collected to service AI applications in medicine and health to be used in harmful ways or for nefarious purposes. It has to be balanced against the fact that safeguards are inadequate.”