What COVID-19 has taught us about the adoption of AI in medicine and healthcare

COVID-19 has accelerated the adoption and scaling of AI. AI is becoming a vital necessity for medicine and healthcare, but much remains to be done to make it more effective, particularly regarding access to data and data sharing by companies and patients.

Effective AI needs high quality input data, but much data is typically siloed across an ecosystem of players, and there is widespread resistance to sharing it with others. More data is available than we might think, but much of it can be described as a swamp of disorganised information. The aim should be to create a data platform, integrating the right data to solve the right problem. In the context of COVID-19 for example, the impact needs to be measured across multiple dimensions: not just health but also in terms of the social and economic impact, such as insights into the effect of government restrictions on infection levels.

There is also a general fear among individuals about sharing personal data, with concerns about security and privacy, and a lack of trust.  Patients are unwilling to consent to sharing when they do not understand the purpose. A recent study conducted by Deloitte found that if physicians explain the reasons in detail and in an understandable fashion, the rate of consent increases from 30% to 90%.

Research also continues into the relationship between mobility (and social distancing) and the spread of infection and outbreak of disease. Vast amounts of data can be extracted from social media platforms and telecommunication operators, but efficacy and adoption depends on the quality of the data, in a usable form and available in real time.