Artificial Intelligence (AI) is everywhere. At least five articles in my inbox this morning had an AI hook. And whether people are singing its praises ("AI will improve your speed, efficiency, and quality of work!") or signaling its pitfalls ("Forget privacy as we [used to] know it!"), the fact is that AI is part of our everyday life – even if you aren't on social media or riding in a driverless car.

In health care, the upside sell of AI is enticing. Data is - and always has been - everywhere in health care. In straightforward applications, AI promises to help providers work through data faster and more accurately to improve patient outcomes, or to spend less time recording data to spend more time with direct patient care. The ability to aggregate and analyze swaths of data ranging from pharmaceutical compound reactions and clinical trial outcomes to patient therapy responses and adverse events is not new, but as AI capabilities improve and expand, the health care applications are exciting.

Our ML Strategies colleague, Christian Tomatsu Fjeld, spent years as senior staff on the Senate Commerce Committee. Since AI falls under the Commerce Committee's jurisdiction, he has deep experience considering the federal policy issues posed by AI. Recently, Christian and other experts joined a panel discussion hosted by the San Francisco Business Times where they discussed AI's economic impact and what businesses should do to effectively leverage AI, as well as public policy and legal considerations for AI's increasing presence in our daily lives. Their insights apply across all industries, but I found the panel's discussion useful in considering some of the same questions around the promises and limitations of AI in healthcare.

For example:

  • How will AI impact the workforce? Some have warned that AI will replace jobs, while others point to the 'elevation' of workers' day-to-day tasks. In health care, will AI complete more and more basic tasks so caregivers can be more efficient with their time? A huge portion of the health care system deals with data entry, submission, analysis and review – just think about claims processing. AI already brings efficiencies to that process, but as the panelists discuss, a big limitation to AI is "dirty data" – an issue that can plague EHR systems. So how can we use AI to improve the accuracy and usefulness of the data we already have?
  • Will the beneficial effects of AI be equitably distributed? Will AI improvements propel larger, capital-rich systems and for-profit health care companies forward while leaner local and regional providers are left behind? Or will there be more democratic AI solutions that are available and affordable to all?
  • Does our existing regulatory framework support the growing impact AI has on our day-to-day lives? Many are rightfully concerned about privacy and security implications of the broad dissemination and use of data in AI. The Senate and House are both considering bills that may supplement existing privacy protections, and my colleagues published a recent viewpoint looking back at the history of HIPAA and speculating about what we might expect from Congress.

These questions are just a few that the panelists considered in their discussion, and many of the answers will lead to more questions. But that's one promise of AI – to use its increased power and sophistication to tackle questions new and old.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.