Artificial intelligence companies are taking on the health industry with big results for consumers. Here are the pros, cons and potential business opportunities.

Try to conceptualize all the medical knowledge in the world. How many books would it fill? How many topics would it cover? How many years would it take to learn in its entirety?

Now, imagine it doubling.

In 1950, researchers predicted it would take about 50 years for all available medical knowledge to double. But in 2020, estimates peg it at just 73 days.

Enter artificial intelligence (AI). It’s not humanly possible for medical professionals to keep up with the influx of constant new information about health conditions, treatments and medical technology. That’s why the healthcare industry is a real growth opportunity for AI and machine learning — and the companies creating smart tools. Healthcare’s artificial intelligence market should increase elevenfold between 2014 and 2021, according to research from Accenture — that’s $600 million to $6.6 billion.

The difference between AI and machine learning

First, a quick detour to define AI and machine learning. Remember your elementary school teacher telling you that all squares are rectangles, but not all rectangles are squares? AI and machine learning have a similar relationship: Machine learning is one application of AI, and it operates on the premise that machines can use extensive amounts of data to learn for themselves and eventually recognize more of the unknown. AI — the advancement of computer systems to perform tasks usually limited to humans — is a broader concept that can be used for other applications besides machine learning. This has constantly developing implications for both consumer health and the healthcare industry as a whole.

How AI is transforming healthcare today

Machine learning easily integrates with aspects of the healthcare industry that have extensive data sets and past examples readily available. The software can learn patterns to help detect tumors, recommend diagnoses and even predict details of a patient’s hospital stay. For example, Google recently introduced a new algorithm for a certain type of software called a neural network. The company’s new algorithm aims to predict hospital patient outcomes such as length of stay, chances of readmission and odds of death.

But it’s not all about poring over data sets: AI tools can also serve as consultants for medical professionals. Take IBM’s Watson supercomputer, for example. Introduced in 2010, it combines AI with analytical software to answer questions — and the company says it’s processed more than 115,000 patients and consumers. Researchers are constantly publishing new information and breakthroughs on certain conditions, so many doctors use Watson to stay up to date on the latest findings, says Alan Smeaton, professor of computing at Dublin City University. The software is in use in hospitals around the country to help doctors gauge the impact of certain symptoms, come to a diagnosis and make decisions.

As for how doctors feel about the new tools available? It all depends on whether something improves their work with respect to their ultimate authority, says Robert Kaul, CEO of Cloud DX, a medical device technology company. As long as medical professionals have the final say on a diagnosis, “we found clinicians are actually excited about the concept,” he says. But AI and machine learning tools in healthcare have a high bar to meet. Although human error is one of the principal reasons AI is being introduced to healthcare in the first place, one wrong turn with an AI tool could mean setbacks in the public eye and the industry itself — especially when it comes to potential bias.

“We know people of color are much less likely to get good healthcare, as are low income people in rural areas,” says American civil rights activist Maya Wiley. “Is the data essentially biasing outcomes because of insufficient data? And what does it mean to collect that data?”

Opportunities for entrepreneurs

IBM’s Watson is a more-or-less unique industry tool with minimal competition, Smeaton says. Entrepreneurs looking to break into AI should know that 30 percent of businesses are predicted to incorporate AI before 2019. That’s up from just 13 percent last year, according to Spiceworks, an information technology company. And a foray into AI for healthcare could mean even more growth. That’s because in the modern world, “data is currency,” Kaul says. In an industry where every individual’s health data could fill 300 million books during their lifetime, there’s plenty of financial opportunity in making sense of that information.

In that same vein, some entrepreneurs are incorporating AI into wearable health devices to save consumers time and money. These tools could democratize some aspects of healthcare that used to require a hospital visit — think assessments for diabetes, Parkinson’s disease, multiple sclerosis (MS) and more. One example is Sensoria Fitness, a company that creates AI wearables for runners. Currently, postural sway (or likelihood of falling) is measured by expensive in-hospital machinery. But research published by the Institute of Electrical and Electronics Engineers (IEEE) shows that Sensoria’s “smart sock” embedded with sensors could be a viable substitute to help predict balance impairment in a low-cost way.

Similarly, some AI wearables allow individuals to share real-time health data with doctors. Qardio, an AI health company, introduced an ECG/EKG wearable that promises a higher diagnostic yield — or a better chance of collecting the information necessary for a diagnosis — in a much shorter amount of time. That, in turn, could lead to shelling out less money for health insurance premiums.

Why this field could be the future

Over the past decade, the healthcare industry has focused on digitizing medical records. Now, the problem is managing those massive data sets — so any infrastructure that can make that possible should be in high demand over the coming years.

“[Using AI], we can take a run at solving some of the hard and some of the previously impossible-to-solve problems in healthcare and in life sciences in general,” says Zachary Bogue, co-founder and co-managing partner of DCVC, a deep tech and AI venture fund. He says he believes the value of the field will “increase materially — and likely increase exponentially” over the next five years.

Bogue says he sees entrepreneurs in his company’s office every week claiming to have the world’s best AI algorithms, but he asks them, “How long will that lead last?” It’s important to have something to carry you past the initial idea. Bogue’s fund invests in entrepreneurs with three elements, he says: world-class AI algorithms, deep industry experience and a proprietary data set (or access to one). He says it’s a “huge opportunity right now” because the healthcare industry has many proprietary or semi-proprietary data sets that entrepreneurs can access through partnering with different organizations.

One more tip for entrepreneurs looking to break into the space? Founders should be both experts in AI and in healthcare — and if you’re one or the other, it’s a good idea to seek out a co-founder whose experience complements yours.

“Healthcare is such a huge, lumbering industry and [so] incredibly complicated,” Bogue says, “that a plucky young AI expert is going to have a very hard time solving any hard problems without … aligning him- or herself with someone with deep industry experience, and vice versa.”