A revolution is on the horizon -- artificial intelligence has the potential to transform healthcare and disrupt the field of medicine in unimaginable ways. Showing remarkable progress in diagnostics, data analysis, and precision, artificial intelligence has already been applied in various fields ranging from patient triage to cancer detection. Complex and challenging for stakeholders, it remains clear that artificial intelligence has metamorphosed various fields, with the potential to improve patient care and quality of life. Faster diagnoses, fewer errors, and functional autonomy: what does this mean for the future? 

 

Whilst it holds excellent potential for improving disease diagnosis, treatment selection, and clinical laboratory testing, increasing accuracy, reducing costs, and saving time, there are imperative obstacles that must be addressed (Harvard). The first challenge arises in the mere integration of artificial intelligence into medical systems. These technologies are continuously learning from new data, and as such, patient data must be continuously absorbed into the systems. A recent study suggested that 85% of consumers are concerned about data privacy (Data Grail), in which it is expected that patients will question the transparency of the usage of their personal data and its safety. Subsequently, a predominant issue of artificial intelligence is its bias: when creating datasets for a model's training, representation bias arises if those datasets do a poor job of representing the users of the model. What will be the repercussions if humanity’s bias is systematised? 

 

The US Department of Health and Human services defines health equity as the “absence of avoidable disparities or differences among socioeconomic and demographic groups or geographic areas in health status and health outcomes such as disease, disability, or mortality.” However, despite aspirations towards the positive impact of artificial intelligence mitigating humanity’s biases, numerous research have demonstrated instances of how the application of AI in healthcare have exacerbated, rather than reduces, health inequities. This is particularly true in cases where AI-using systems deviate from quickly developing evidence-based standards or when they may have been intended for non-commercial usage but end up being employed more widely. 

 

Any aspect of the healthcare system that contains bias has the potential to have varying effects on different populations, and historically, this has meant that underrepresented, underserved, and under-resourced populations have had worse health outcomes. An analysis of a popular AI system revealed that, despite requiring higher acuity treatment, sicker black patients received similar care to healthier white patients -- a result of this inappropriate usage of the AI system. Based on these ethical frameworks and the ongoing creation of "metrics for ethics," bias is measured as the variation in how a healthcare procedure affects a specific group. 

 

But artificial intelligence learns from the data set given -- given by humans. Whilst bias remains prevalent in artificial intelligence, it is humans who offer healthcare that exhibits this discrimination. A recent study (Char et al) revealed that provider bias, in which notes kept by doctors, recorded Black patients' symptoms and indicators in a more disparaging way, leading to the prevalence of inequity in the medical field; this prejudice goes further than just race too, impacting various marginalised and underrepresented communities and groups globally. 

 

But the artificial intelligence research about bias in healthcare is remarkably in the early stages. This is because these studies have numerous limitations that require future investigation. What remains clear is that more work must be done -- having a diverse and inclusive sampling, method triangulation, and peer review, must be implemented for artificial intelligence to be used at its full potential. The dawn of a new era is upon us all: of apocalyptic discrimination or amplified ingenuity. The choice is ours. 

 

 

Works Cited: 

Abràmoff, Michael D., et al. “Considerations for Addressing Bias in Artificial Intelligence for Health Equity.” Npj Digital Medicine, vol. 6, no. 1, Nature Portfolio, Sept. 2023, https://doi.org/10.1038/s41746-023-00913-9. Accessed 3 Feb. 2024.

“Artificial Intelligence in Medicine | IBM.” Ibm.com, 2024, www.ibm.com/topics/artificial-intelligence-medicine#:~:text=How%20is%20artificial%20intelligence%20used,health%20outcomes%20and%20patient%20experiences. Accessed 3 Feb. 2024.

“Artificial Intelligence in Medicine | the Top 4 Applications.” Datarevenue.com, 2022, www.datarevenue.com/en-blog/artificial-intelligence-in-medicine. Accessed 3 Feb. 2024.
‌Basu, Kanadpriya, et al. “Artificial Intelligence: How Is It Changing Medical Sciences and Its Future?” Indian Journal of Dermatology, vol. 65, no. 5, Medknow, Jan. 2020, pp. 365–65, https://doi.org/10.4103/ijd.ijd_421_20. Accessed 3 Feb. 2024.

Briganti, Giovanni, and Olivier Le Moine. “Artificial Intelligence in Medicine: Today and Tomorrow.” Frontiers in Medicine, vol. 7, Frontiers Media, Feb. 2020, https://doi.org/10.3389/fmed.2020.00027. Accessed 3 Feb. 2024.

Dr. Varsha P．S. How Can We Manage Biases in Artificial Intelligence Systems – a Systematic Literature Review. no. 1, Apr. 2023, pp. 100165–65, https://doi.org/10.1016/j.jjimei.2023.100165. Accessed 3 Feb. 2024.‌

Gast, Kelsey. “Artificial Intelligence – a Danger to Patient Privacy? | LogRhythm.” LogRhythm, SIEM Platform & Security Operations Center Services | LogRhythm, 28 Aug. 2023, logrhythm.com/blog/risks-of-artificial-intelligence-in-healthcare/#:~:text=As%20artificial%20intelligence%20infuses%20its,their%20data%20is%20overall%20safe. Accessed 3 Feb. 2024.

Help Net Security. “Consumers Take Data Control into Their Own Hands amid Rising Privacy Concerns - Help Net Security.” Help Net Security, 11 Apr. 2023, www.helpnetsecurity.com/2023/04/11/personal-data-privacy-concerns/#:~:text=Data%20privacy%20concerns&text=As%20such%2C%20people%20are%20actively,data%20and%20for%20what%20purpose. Accessed 3 Feb. 2024.
“How Artificial Intelligence Is Disrupting Medicine and What It Means for Physicians.” Harvard.edu, 13 Apr. 2023, postgraduateeducation.hms.harvard.edu/trends-medicine/how-artificial-intelligence-disrupting-medicine-what-means-physicians. Accessed 3 Feb. 2024.
“Shedding Light on Healthcare Algorithmic and Artificial Intelligence Bias.” Office of Minority Health, 2023, minorityhealth.hhs.gov/news/shedding-light-healthcare-algorithmic-and-artificial-intelligence-bias#:~:text=Healthcare%20algorithms%20and%20AI%20bias,used%20to%20train%20computer%20programs. Accessed 3 Feb. 2024.