Although artificial intelligence has been touted as the next big thing in healthcare, its adoption within health systems is very slow. This may be due to algorithmic limitations, data access issues, regulatory problems and misaligned intentions, according to a March 9 Brookings Institute report.
Books have been written on health AI and many academic and industry conferences have created special AI showcases, indicating industry excitement about the promises of AI. However, in practice, the picture looks different. When looking at which roles require AI skills in job board postings across industries, healthcare sits near the bottom, with only 1 in 1,850 jobs requiring AI skills. This means just 0.05 percent of healthcare and social assistance jobs require AI skills.
Here are four main reasons why healthcare hasn't jumped on the AI trend just yet, according to the Brookings Institute:
- Some AI algorithms work in ways that are difficult to understand, in turn making it a challenge to understand how they produced a certain result. In healthcare specifically, when medical professionals are held accountable for decisions made by algorithms, they need to fully understand how the result was generated.
- High quality data sets are integral to high-performing AI algorithms. However, given issues with EHR compatibility and the difficult and time-consuming process of collecting patient data, many hospitals do not have access to such rich data sets.
- Regulation in healthcare is more cautious than other industries, creating barriers for AI implementation. Patient data privacy and security regulations, device approval and liability concerns all make AI challenging to fully integrate into the industry.
- The level of implementation of AI at health systems also depends on the hospital's leadership, and the intentions of each decision-maker. Some physician-led hospitals may not favor AI for fear it will replace them, and other administration-led hospitals are more open to adoption of AI in non-clinical roles.