Health Science Reviews

By Thato Makena

The human population is at the highest it has ever been, and potentially the highest it will ever be. Globally, surveys estimate that around 50-70% of healthcare professionals suffer from burnout, stress from increased workloads and an unhealthy work-life imbalance. This is concerning because people’s lives are in the hands of the very same people.
Artificial intelligence is an increasingly popular tool that simplifies and enhances the way we live and work, but despite the concerning stats above and the increase in investments for AI in healthcare and medicine, there are still so many barriers and controversies preventing its implementation in real life.

By conducting a systematic review using search terms like “AI/Machine learning/Deep learning” combined with “Health/Medicine” on platforms like PubMed and Google Scholar, it is easy to gauge the multiple ways in which AI can potentially revolutionise and augment the healthcare industry. Such a review also highlighted three key limitations and concerns surrounding using AI in this industry, and it is those points that keep it in a preclinical space and prevent its
implementation in real-life settings.

Technical challenges, such as data handling and management, are a major issue. Healthcare deals with sensitive and private information, which is an issue because AI models rely on very large datasets. Incentivising and promoting the shift from prioritising individual treatments to overall patient outcomes can promote data sharing, which will result in better and more reliable/accurate AI models.

Ethical concerns like patient privacy, consent and accountability are also major roadblocks. Prioritising the anonymisation of patient data as well as informed consent from patients is key. Patients should be aware that their data is being used to train AI models. Accountability is also a concern; if a patient is harmed because of decisions made using these technologies, who bears the responsibility?

Lastly, social concerns surrounding AI are also an issue. There is a growing concern that AI will take over people’s jobs, which is a massive misunderstanding and overestimation/overexpectation of what AI can do. Involving all stakeholders in its implementation can help lessen fears around AI, and by using frameworks and visualisation tools to explain why the technology made the decision it made can build trust in the technology.

Overall, it is understandable why lots of AI-based technologies are stuck in the pre-clinical and experimental stages of implementation. Co-operation between physicians, scientists, legal and ethical bodies, as well as the public, is needed for this to work. Prioritising patient care over the excitement of new and ground-breaking technology is also very important. Slow, careful and stepwise adoption of AI into the healthcare industry is just as important.

Addressing all concerns and misunderstandings surrounding AI is crucial because, at the end of the day its purpose is to re-engineer jobs and allow healthcare workers to focus on the more innately human, emotional and unpredictable aspects of medicine.

References:

1. Peng, Z. and Ren, X. (2024) ‘Application and development of artificial intelligence-based
   Medical Imaging Diagnostic Assistance System’, International Journal of Biology and Life
   Sciences, 6(1), pp. 39–43. doi:10.54097/sb3m1m17.
2. Aung, Y.Y., Wong, D.C. and Ting, D.S. (2021) ‘The promise of Artificial Intelligence: A
   review of the opportunities and challenges of artificial intelligence in Healthcare’, British
   Medical Bulletin, 139(1), pp. 4–15. doi:10.1093/bmb/ldab016.