AI in Healthcare: Saving Lives and Money—But Change Takes Time

Picture entering your physician's clinic feeling unwell—instead of sifting through stacks of your medical past or waiting for lab work that spans several days, your doctor quickly gathers information from your electronic health files, DNA makeup, and fitness trackers to figure out what's causing your illness.

Subscribe to our newsletter for the most recent science and technology news developments.

Such fast diagnostic methods represent one of the major hopes for artificial intelligence within the healthcare field. Advocates of this technology believe that in the next few decades, AI could help save hundreds of thousands of lives. even millions of lives .

Furthermore, a 2023 research indicated that if the healthcare sector greatly expanded its adoption of artificial intelligence, as much as Approximately $360 billion per year might be saved A yearly saving of US$360 billion could be achieved Annual savings of up to US$360 billion may be possible Up to US$360 billion each year could be conserved Savings amounting to US$360 billion every year might be realized .

However, despite artificial intelligence becoming almost everywhere—ranging from mobile devices to chat applications to autonomous vehicles—the effect of AI on healthcare remains fairly limited thus far.

A 2024 survey conducted by the American Medical Association revealed that 66% of U.S. doctors have utilized artificial intelligence tools in some form, an increase from 38% in 2023. However, much of this usage was for administrative or low-risk support And even though 43% of U.S. healthcare institutions introduced or broadened their use of AI in 2024, numerous deployments are still exploratory especially with regard to healthcare choices and health assessments.

I'm a professor and researcher Who specializes in artificial intelligence and healthcare analysis. I will attempt to clarify why the expansion of AI will occur slowly, and how technological constraints and moral issues hinder the broad implementation of AI within the healthcare sector.

Inaccurate diagnoses, racial bias

Artificial intelligence excels at finding patterns in large sets of data. In medicine, these patterns could signal early signs of disease that a human physician might overlook—or indicate the best treatment option, based on how other patients with similar symptoms and backgrounds responded. Ultimately, this will lead to faster, more accurate diagnoses and more personalized care.

AI can also assist hospitals in operating more effectively By examining processes, forecasting staff requirements, and organizing surgical schedules to ensure optimal use of critical assets like operating rooms. Through simplifying tasks that typically require extensive manual work, AI enables healthcare workers to concentrate more on hands-on patient treatment.

But despite its strength, AI can make mistakes Although these systems are developed using information from actual patients, they may have difficulty when faced with rare situations, or when the data does not exactly align with the current patient.

Consequently, artificial intelligence does not consistently provide a correct diagnosis. This issue is referred to as algorithmic drift —when artificial intelligence systems function effectively in controlled environments but experience reduced precision in actual scenarios.

Another concern is racial and ethnic discrimination. If data include bias since it lacks sufficient representation from specific racial or ethnic communities, AI could produce unreliable suggestions for these individuals, resulting in incorrect diagnoses. Certain findings indicate this has already happened .

Data-sharing concerns, unrealistic expectations

Healthcare systems are highly intricate and complicated. The idea of incorporating artificial intelligence integrating into current processes can be challenging Introducing a new technology such as AI can change everyday activities. Employees will require additional instruction to utilize AI systems efficiently. Numerous hospitals, clinics, and medical practices lack the available time, staff, funds, or desire to adopt AI.

Additionally, numerous advanced AI systems function as mysterious "black boxes." They generate suggestions, yet even their creators may find it challenging to completely clarify the process. This lack of transparency conflicts with the requirements of healthcare, where choices require clear reasoning.

However, programmers frequently hesitate to reveal their private algorithms or information providers , both to safeguard intellectual property and due to the difficulty in simplifying such intricacies. The absence of clarity fuels doubt among professionals, leading to delays in regulation and diminishing confidence in AI results. Numerous specialists contend that transparency is more than merely an ethical consideration but rather a necessary step for implementation an essential requirement for taking action a vital consideration for acceptance a crucial factor for implementation an indispensable element for application a fundamental aspect of embracing change a key component for execution an important prerequisite for incorporation a mandatory condition for utilization a significant criterion for integration in health care settings.

There are also privacy concerns ; data sharing could threaten patient confidentiality 1. Medical AI systems typically need large volumes of patient data to train algorithms or generate forecasts. Improper management can lead to the disclosure of confidential health details, either via data leaks or unauthorized access to medical files. 2. In order to develop effective algorithms or produce accurate predictions, medical artificial intelligence relies heavily on extensive patient datasets. Without proper safeguards, this technology may inadvertently reveal private health information, such as through security incidents or improper handling of records. 3. Training machine learning models or making clinical predictions with medical AI usually requires vast quantities of personal health data. When not managed carefully, these systems risk exposing delicate medical information due to hacking attempts or misuse of patient data. 4. For medical AI applications like algorithm training or diagnostic forecasting, significant amounts of patient data are necessary. Failure to manage this data appropriately might result in privacy violations, including exposure of protected health info through breaches or unapproved usage of medical records.

For example, a healthcare professional utilizing a cloud-hosted artificial intelligence tool to help write a medical record should make sure that no person without permission can view the patient's information. U.S. laws for example, the HIPAA legislation implement stringent regulations for the exchange of health information, indicating that AI creators must establish strong protective measures.

Concerns about privacy also affect patient confidence: When individuals worry that their health information could be improperly used by an algorithm, they might become more hesitant or completely decline care supported by artificial intelligence.

The great hope of artificial intelligence is a significant obstacle in its own right 1. The expectations are very high. Artificial intelligence is frequently depicted as a miraculous remedy capable of diagnosing all illnesses and transforming the healthcare sector instantly. Such unrealistic beliefs typically result in frustration. AI might not quickly fulfill its potential. 2. There are significant hopes surrounding AI. It’s commonly seen as an extraordinary tool that can identify every illness and completely transform medical services within days. These overly optimistic views usually end up causing letdowns. AI may take time before meeting these expectations. 3. High anticipations exist. Many people view artificial intelligence as a wonder cure-all able to detect any condition and rapidly reshape the entire healthcare system. Misguided notions such as this tend to cause disillusionment. AI could require more time than expected to meet those claims.

Lastly, building an effective AI system requires significant experimentation and adjustment. AI systems need to undergo thorough evaluation to ensure their safety and effectiveness It requires many years, and even once a system has been authorized, modifications might still be necessary as it processes different kinds of data and faces various real-life scenarios.

Incremental change

Currently, medical facilities are quickly embracing AI scribes that monitor conversations during patient consultations and generate clinical records automatically, minimizing administrative tasks and allowing doctors to dedicate more time to their patients. Studies indicate that more than 20% of physicians currently utilize AI for documenting patient care during hospital stays or summarizing treatment upon release AI is increasingly serving as an unobtrusive element in administrative tasks. Hospitals utilize AI-powered chatbots to manage appointment bookings, address typical patient inquiries, and provide instant language translation.

AI has practical applications in clinical settings, although they remain restricted. In certain hospitals, AI serves as an additional tool for radiologists. searching for initial indicators of illness trying to identify symptoms at an early stage seeking warning signals of health problems looking for preliminary signs of medical conditions hoping to detect issues before they worsen investigating early manifestations of diseases wanting to recognize first clues of sickness attempting to spot early warnings of ill health However, doctors remain hesitant to transfer decision-making to computers; only approximately 12% of them do so at present. depend on artificial intelligence for medical assistance use AI technology for diagnosis support turn to machine learning for health evaluation leverage automated systems for clinical guidance employ intelligent algorithms for disease detection utilize digital tools for diagnostic purposes seek computer-based solutions for medical analysis opt for algorithm-driven methods for health assessment consult smart software for condition identification apply advanced computing for accurate diagnoses .

It goes without saying that the shift of healthcare toward artificial intelligence will occur gradually. New technologies require time to develop, and current demands within healthcare continue to take precedence over future benefits. Meanwhile, the ability of AI to help countless individuals and reduce massive costs remains on hold.

This piece is reposted from The Conversation Licensed under a Creative Commons license. Read the original article .

Provided by The Conversation

This narrative first appeared on Medical Xpress .