Understanding the Role of Artificial Intelligence in Healthcare

As frequent readers may recall, the last time I reported on artificial intelligence (AI) I was discussing the big question marks surrounding the role of AI in healthcare. 

Well, last week we got a couple of answers, a look at a potential new method Congress may use to legislate AI, and an announcement that might indicate healthcare will pave its own way in what’s being called by Senate Majority Leader Schumer the “AI revolution.”

Senator Schumer held a press conference to officially present his framework for congressional legislation of AI, making good on a promise from April where he announced his plans to work towards refining the guardrails Congress will put on the emerging technology. Entitled the SAFE Innovation framework, the title represents each of the pillars of Senator Schumer’s proposal: security, accountability, foundation, and explainability. These concepts make up the word “safe,” while “innovation” is what Senator Schumer called the “North Star” of all AI policy. 

Security refers to protection from both external security threats as well as protecting the American workforce from AI eliminating their jobs.  Accountability means making sure that AI isn’t used in illegal or harmful ways—particularly when it comes to stealing intellectual property. 

What areas could AI be used, but shouldn’t? Foundations of our country include democracy, justice, and our electoral process-and AI cannot be allowed to be programmed to undermine them.  And finally, “explainability” means being able to determine why AI chose one answer over another to ensure accuracy. 

Senator Schumer also spoke to a new way Congress might be operating when it comes to legislating AI because it’s such a new issue.  He stated that this new legislative process will involve what was called insight forums with the quote “top minds in AI” rather than traditional congressional hearings with questioning. 

Schumer believes regular hearings would simply be too slow because of the rapidly changing environment of AI and if they took that approach, the actions by Congress would be obsolete by the time they passed anything. These new forums could allow legislators to simply learn-learn about the great potential and the great risk of AI.  And then take that back to the drawing board to craft legislation.

The Senate already has three scheduled briefings on AI to be held throughout 2023: the first would be a general overview of AI, the second on how to achieve American leadership on AI, and the third on general defense and intelligence concerns.  Senator Schumer promised in his speech that a legislative package would not take years-although “weeks” was too optimistic and it should be noted that the AI briefings are scheduled into early winter.

You might be noticing, though, that healthcare isn’t specifically represented or mentioned in these discussions on creating AI boundaries yet.  However, two academics from Stanford hope to lead the way on AI policy in the industry.

The Dean of Stanford School of Medicine and a director of the Stanford Institute for Human-Centered Artificial Intelligence just launched an initiative that aims to determine how to use AI in health and medicine responsibly.  The goals are to use AI to enhance clinical care outcomes, accelerate research on the biggest issues facing the industry, and assist in educating patients, providers, and researchers.  The initiative aims to do this by developing a go-to platform for AI resources and defining AI standards and safeguards for use in healthcare.

Although we have a little more clarity on the response by Congress rto AI, it’s equally possible that the healthcare industry might be taking matters into their own hands in at least developing recommendations and best practices for use in clinical settings.