Advances in artificial intelligence have the potential to completely transform healthcare. Innovative applications for AI are changing the way healthcare workers do everything from scheduling an appointment to diagnosing cancer. Current events, including the COVID-19 pandemic and increased awareness of the impact of biased training data sets, are further impacting how the industry uses AI to care for patients.
As the population ages, the US will need a growing number of healthcare professionals. AI is positioned to help fill the shortage of skilled workers, if the industry can address questions surrounding the ethics of AI in healthcare. Physicians promise to care for their patients while following the American Medical Association’s nine Principles of Medical Ethics. How do we hold AI providers working in tandem with human doctors to the same ethical standards? When care is provided by both AI and humans, who (or what) is responsible for errors?
This need for ethical AI begins long before an algorithm plays a part in patient care. Because much of AI-enabled decision making happens within a “black box,” it’s difficult for healthcare providers and their patients to trust these decisions. This is compounded by a lack of transparency in clinical trials and “scarcity of external validation studies.” Machine learning has the potential to impact existing biases in healthcare. Whether this impact will be positive or negative depends on the data used to train AI tools. Without much insight into these data sets and results of randomized controlled trials, gaining the trust of skeptical clinicians and patients is likely to be difficult.
The ability of AI to predict future trends using past data is incredibly valuable in healthcare, an industry constantly innovating to protect patients against well-known chronic illnesses and novel concerns alike. During the COVID-19 pandemic, AI predicted “under-studied proteins associated with SARS-CoV-2, the virus that causes COVID-19.” This work informs research into potential treatments, which is also assisted by machine learning. Public health initiatives rely on AI as well. AI is currently used to map the spread of COVID-19 and detect misinformation about the virus online.
AI can offer personalized care before and after an individual receives treatment from their healthcare provider. These types of interactions, such as booking appointments and gathering information from patients before an appointment, are labor-intensive when completed by humans. Allowing AI to take on some of the patient-provider communication makes these processes more convenient for both parties. Personalized preventative care, such as recommendations based on health history and other factors, is likely only possible at scale with the help of AI.
Diagnosing an illness or other health issue with diagnostic tests requires incredible attention to detail. AI is well-suited for this type of task, due to its “higher accuracy and lower turnaround time and cost in comparison to non-AI techniques.” After diagnosis, AI may also be able to further personalize treatment plans, including pharmacologic treatment.
Just like other industries, the impact of AI in healthcare is felt beyond frontend interactions. AI is used in human resources, finance, IT, and other administrative tasks. This allows practitioners and administrators alike to focus their limited resources on patient care. AI’s data management excellence may also play a role in managing and monitoring clinical trials.