You are here
Diagnosing a business case for AI in healthcare
Not a day goes by these days without hearing about how artificial intelligence (AI) is set to reshape industries, professions and everyday lives. And while AI is certainly not a magic bullet that will solve all problems with a snap of our fingers, its transformative power, when done right, simply cannot be denied.
The projected numbers are staggering. By Allied Market Research’s estimates, the global AI market is set to pull in US$169.41 billion by 2025 at 55.6% compound annual growth rate. With various industries putting their collective weight behind the advancement of AI, how will it impact the future of work?
To find out more about AI’s use cases, Michael Page recently invited Dr Veerachat Petpisit, CEO of Healthcare Ventures, BridgeAsia, to share how AI works in the healthcare context. He addressed how various industries can harness AI’s capabilities, as well as potential pitfalls that leaders should watch out for. The recent event, held at Michael Page Singapore, attended by around 40 leaders from various industries.
Rise of the machines
For the uninitiated, AI is essentially a collection of technologies that solve problems and perform tasks without explicit human guidance. Automation and machine learning are at the core of how it works. As Dr Petpisit explains, while humans can generally handle decisions involving up to 10 factors at any a time, tasks that involve millions or even billions of data points can become overwhelming very quickly.
But not for AI. To put things into perspective, he shared that sequencing a single human genome involves about 100GB of data, while a driverless car collects about 2TB of data per session. An average flight in a modern aircraft? You’re looking at up to 8TB of data. Processing this volume of data is something humans simply cannot take on. And it’s not about machines ‘replacing’ human counterparts, either. Rather, AI is there to do complex and repetitive tasks that humans cannot, which in turn lets us to focus on higher-value work.
Part of the reason why AI is coming into its own in recent years, aside from more powerful computers and faster communication speed, is the rise of collaborative business models. “A good example of this is cloud computing,” Dr Petpisit explains. “The cloud is basically a place where somebody with lots of computers in the same place has them up for rent. Many companies now have a very low start-up cost because they can rent these machines, they can rent computing power, they can rent storage. It’s not going to impact your investments and cost.”
In the pharmaceutical industry, Dr Petpisit explained that even small companies have the capacity to discover new drugs. The ability to share computational power too, is a big part of why AI has gone mainstream.
AI in action
In terms of AI’s use cases in healthcare, Dr Petpisit cites a few examples. AI can now read chest X-rays and spot up to 14 diseases, such as pneumonia, edema, nodules, atelectasis, at 80% accuracy, or about as accurately as a real-life doctor. This means doctors can provide faster service, diagnose issues and spend more time focusing on the patients themselves.
Dr Petpisit also mentioned that while it’s not uncommon for even experienced dentists to miss up to 40% of cavities, a similar image recognition technology might also reduce or even remove instances of such omissions altogether. And this is just skimming the surface of what AI has the potential to do for healthcare. Beyond the examples provided by Dr Petpisit, AI-powered voice-to-text transcriptions can help automate tedious, repetitive non-patient care activities like writing chart notes, prescribing medications and ordering tests.
By Forbes’ estimates, this alone could result in US$18 billion in savings. Virtual nursing assistants meantime, could save the healthcare industry US$20 billion every year. Since these AI assistants are available around the clock, they can answer questions, monitor patients and provide quick answers anytime, anywhere. This allows for a more consistent and reliable communication between patients and their care providers – and could potentially reduce hospital readmissions or unnecessary visits.
Words of caution
Beyond healthcare, there are plenty of real-world applications for AI, and it really differs from one industry to the next. With that said, Dr Petpisit is quick to warn commercial attendees that this is no golden ticket. “AI is not magic. It is not going to suddenly solve all your problems just because you decide to implement it tomorrow.”
For companies looking to invest in AI, Dr Petpisit said that there are three criteria that need to be satisfied. The first is the inherent business value. As with any investment a company makes, be it a product, service or talent, it has to first serve business interests. Just because other companies are jumping on a bandwagon does not mean that you necessarily need to partake.
The second is what Dr Petpisit refers to as current accessible assets. “By assets, I’m talking about whether the company has good data, the right talent – like data scientists – as well as the budget to carry out AI implementation,” he explains. “A lot of companies say that they have a lot of data, but the truth is that a lot of it tends to be missing, wrong, inappropriate or duplicates,” he says.
Finally, it all comes down to an organisation’s long-term support. Dr Petpisit observes that it is not uncommon for companies to “run out of momentum” when it comes to AI implementation. “Some people want to use AI to reduce cost: and that’s fine. However, the truth is that AI does not offer ROI immediately. Things don’t happen right away,” he warns. On top of that, he said that there are external factors as well, from government policies right down to the quality of data scientists. “So for AI to work, you really need that long-term support.”