The atmosphere of AIMed North America 2018 is so cordial that it makes the artificial intelligence (AI) in medicine industry looks competitive and probably, a little hostile. Chief intelligence and innovation officer of Children’s Hospital of Orange Country (CHOC) and founder of AIMed, Dr. Anthony Chang compared the development of AI in medicine between US, UK and China in a chart during a panel discussion took place on 13 December 2018.
There are five categories on the chart: health data, access need, manpower, academics and government. Both UK and China had each scored a star on health data as Dr. Chang described US still “needs more work” in the area. Indeed, US physicians had expressed their discomfort towards electronic medical records (EMRs) and its unstructured nature make it hard to extract meaningful data. As such, US also misses a star in access need.
In terms of manpower, US has obtained a star, indicating the sector has booming workforce as shown in the number of startups. Dr. Chang believed China will catch up in five years’ time as she invests more money in AI as compared to US. However, China’s interest is not specific to AI in healthcare, unlike US and UK, her strength lies in robotics and facial recognition. Nevertheless, US, UK and China had similar capital in AI research.
Mismatch between demand and reality
The competition between continents may be propelling AI medicine. Yet, “AI provides global health solutions, but we are not doing the basics very well. We are nowhere close to what we want to see in a life time” Dr. Chang said.
On the same panel, Dr. Yu-Chuan (Jack) Li, professor, college of medical science and technology, Taipei medical university said there are many low-hanging fruits for AI in healthcare. For examples, medical error being the third most common cause of death. There is a general poor or inconsistent quality of provided care. Medicine tends to adhere towards average and adopts a “one size fits all” approach and the absent of actionable disease prevention.
All these are opportunities which AI can act on but none of them is duly executed at the moment. Not even when the competition is so furious, creating what is known as the “traffic jam of AI medicine”.
The hinder and solutions
To add to the challenge, commercialization of technology is depleting our access to data, an important component to AI design. “We should make data bases open access. Data shouldn’t be treated as commodity; the money should come from constantly deploying and improving of algorithm” said Dr. Leo Anthony Celi, associate professor of medicine (part-time), Harvard medical school and principal research scientist of Massachusetts Institute of Technology.
Other speakers stressed on the importance of collaboration and long-term development, rather than short-term benefit. “AI allows us to reach out to the community in need of medical service” said Karen Cross, co-founder and chief executive officer of Mimosa Diagnostic. “From engineers to policymakers, it’s important to put the right person in the right team; to promote support and not individuals’ interests,” Cross added.
“Try to reach out to three main stakeholders: academia, industrial and government. For AI medicine to move forward, we need to integrate all these three” said Dr. Celi.
A science writer with data background and an interest in current affair, culture and arts; a no-med from an (almost) all-med family. Follow on Twitter.