The Challenges of AI in the Future of Healthcare

Photo by Owen Beard on Unsplash

With the growing development and sophistication of Artificial Intelligence, AI has the potential to take over many jobs previously performed by humans. The healthcare industry specifically has begun to utilize AI’s advanced reasoning, perception, and learning abilities, and the industry’s shift towards AI is predicted to expand. Developers and potential investors in AI recognize that machines have the potential to make “accurate clinical decision making and amplified R&D capabilities” in healthcare, according to the Harvard Business Review. 

However, before being approved by the U.S. Food and Drug Administration, AI must pass several tests. In response to the growing digital health industry, the FDA launched a Digital Health Innovation Action Plan to ensure safe digital health technologies. In combination with the plan, the FDA also launched the Digital Health Software Precertification Pilot Program in order to “determine the key metrics and performance indicators required for product precertification” (Harvard Business Review). Because AI’s involvement in healthcare is unprecedented, programs like the Digital Health Innovation Action Plan and the Digital Health Software Precertification Pilot Program ensure that AI technologies are effective in their purpose and safe for widespread use. 

While the FDA has not released comprehensive guidelines on AI’s involvement in healthcare, they have begun to allow software “that utilizes machine learning-based algorithms” to aid in clinical decision support (Harvard Business Review). 

Another aspect to consider when using AI in Healthcare are any future updates and security risks. Each time a program needs to receive an update, the update must be approved by the FDA. Therefore, a software developer’s initial programs may be approved by the FDA, but their updates with new generations of software would also need to go through the approval process again. 

Additionally, once AI systems have been implemented, they may face further issues in the future. Much like the autonomous driving industry’s issues with technology making poor human-like decisions such as self-driving technology killing a pedestrian instead of stopping, AI in healthcare faces similar issues. If healthcare workers make inaccurate decisions during the diagnosis of their patient, they could unknowingly comprise the conclusion of the AI system. For example, if a physician makes a misdiagnosis of a patient, the AI system might recommend an incorrect course of treatment. However, despite risk in clinical diagnosis, AI has proved to be very useful in lower-risk scenarios of the healthcare industry such as recording a patient’s history or billing information, which have little room for error.. 

Moreover, even though AI has the potential for widespread use in the healthcare industry and may make clinical decisions in the future, patients have yet to stamp their mark of approval on this innovating technology. According to the  Journal of Consumer Research, patients were less willing to attend appointments where AI was involved in the clinical process. Many patients still view AI systems as detached and lacking the personal touch that only humans can provide. 

Even though AI is informative, accurate, and convenient, it still needs to gain the public’s trust. If developers can convince healthcare consumers that AI can provide a personalized look into patients that takes each patient’s unique characteristics into account, AI has the potential to be successfully implemented across the whole field. Because AI can make the healthcare industry more cost-effective, streamlined, and accurate, it is on the path to become a part of our everyday visits to our local clinics.