Q&A: The effect of Microsoft’s AI for Health initiative
Following the launch of Microsoft’s $40 million five-year AI for Health project, Ian Bolland caught up with Jeffrey Eyestone, healthcare AI advisor at CognitiveScale, to discuss the ambition and potential effects of the initiative.
How ambitious is the investment from Microsoft?
The dollar amount and range of global organisations served, from academia and non-profits to companies with advanced technologies, make this ambitious. Microsoft is reportedly making data scientists available and providing cash grants, so this isn’t just access to Microsoft software and platforms, which could be seen as a bit more self-serving.
The real ambition comes from the range of healthcare use cases that the AI for Health initiative will target: Quest for Discovery, Global Health Insights and Health Equity. AI has the opportunity to make significant gains in these areas, but it is my belief that life-changing results for healthcare as described by these broad AI for Health initiatives will only be achieved through collaboration on this scale, a vision for change that looks beyond profit motive, a desire to address public policy ramifications, investment by government and industry in support of academia and research, and ambition.
You’ve said Microsoft’s willingness to collaborate as important – how important do you think relationship with medical device manufacturers and innovators be?
AI initiatives on the scale of Microsoft's AI for Health will require collaboration on several key levels. Development, orchestration and operationalisation of AI solutions require a level of technical collaboration from data scientists to multiple IT providers if AI is going to deliver value in the form of trusted, production-ready solutions.
Collaboration between research, academia, industry and government looks to be a key aim of AI for Health — but it is vendors like Microsoft who can provide the infrastructure necessary for efficient teamwork all the way to deploying solutions. Microsoft is well-positioned to be an enabler and catalyst for this level of collaboration. One specific and particularly interesting collaboration is AI for Health's Data Collaborative — always key for AI projects — and I'll be curious to see if this develops into a launchpad for innovative device manufacturers and healthcare IT vendors to leverage AI solutions developed under AI for Health.
What can its AI for Health project help to develop?
The three stated initiatives — Quest for Discovery, Global Health Insights and Health Equity — each include numerous compelling AI use cases that will lead to the identification of many possible AI-powered solutions.
For example, Population Health and Social Determinants of Health are areas where AI holds much promise given the reliance on massive amounts of data and challenges with deriving life-changing insights. Microsoft's current AI for Health initiatives include a range of projects that will apply AI to making improvements in the diagnosis and treatment of Sudden Infant Death Syndrome (SIDS), leprosy, diabetic retinopathy and cancer. Microsoft can help by providing a collaborative platform for transforming research and data science into production-ready AI-powered solutions — an area where academia, government and research organisations traditionally struggle.
Is this a game changer for healthtech?
'Game changer' has an element of finality and proof that AI for Health has yet to achieve, so I am not sure I'd use that term just yet. But these ambitious initiatives require more than collaboration and investment. AI for Health needs trusted, proven solutions.
Microsoft's AI for Health is different in that they have the tools and the expertise. More broadly, with its AI for Good investments, Microsoft is showing a desire to take on the large-scale challenges that could be game changers for healthtech and healthcare.
What kind of progress does Microsoft need to make over this time frame?
The list of AI for Health initiatives currently underway is impressive. These projects will require a clear articulation of how to take a proof-of-value project to a production-ready, scalable and trusted solution.
Demonstration of several projects with clear impact on qualitative and quantitative performance metrics will go a long way toward driving value, and I suspect Microsoft and its partners are well-positioned to demonstrate success on a number of projects over this five-year timeframe.