Is AI taking over the world? Or, more specifically, is AI taking over healthcare? With the recent explosion of open AI’s ChatGPT and other AI tools that have gained a lot of attention, both properly and within business, it seems worth having a conversation around what we’re doing with this in healthcare. It’s clear that some of this, even from the software companies, is to pump up stock market valuations, or for bankers to include in acquisition deals. Recruiters and consultants have even suggested sticking “artificial intelligence” into resumes to get through AI screening tools, as that’s something everybody wants to see these days.
Overhyped AI Examples
There are some famous examples of AI in healthcare, one of which is All of AI. They really almost became a perfect example of where AI got overhyped. They had a $4 billion valuation, but the reality is that much of that was Mechanical Turk, with people doing things in the background rather than any true artificial intelligence tools.
Smoke and Mirrors
When we look under the covers of software vendors and applications, we can often see that their claims are more smoke and mirrors or hype than anything else. Something like machine learning is a more specific claim than something broad like “we’re using artificial intelligence.”
Does AI Really Matter?
In the business of healthcare, most problems don’t require AI to solve them. There are situations where only AI can solve a problem due to the complexity of the data, but that’s not the case for most things like analytics around charge reconciliation, credit balances, or provider productivity. Those can be analyzed effectively by an analyst without the use of AI.
Data Challenges
The bigger problem is data. Data is required for AI to both train and enable it to be successful. But we have a massive problem with data in healthcare, particularly on the business side, where data is not easily accessed, transformed, or in a common format. Until we solve this data infrastructure problem, AI is essentially useless. Human
Conclusion: Not Yet
Even if we solve the data challenges, there’s a human capital problem. Healthcare, as an industry, is not data-driven for the business side. Many practice administrators, physicians, CFOs, and mid-level managers lack training in data and analytics, hindering the effective implementation and use of these tools. In conclusion, while AI may be taking over marketing, it has not yet taken over healthcare due to the significant challenges around data and human capital.
