Healthcare AI Will Survive Its Hype Cycle

Posted October 18, 2017
By Julia Zehel


Not long ago, IBM CEO Ginni Rometty took the stage at a major health technology conference to share a bold vision for the future of healthcare. Healthcare AI (or “cognitive computing,” IBM’s preferred term) is poised to usher in a “golden age,” Rometty said. “I believe healthcare could be the leader for the world in showing what it means for an industry to shape an era.”

It’s hardly surprising that Rometty is touting the potential of healthcare AI. Over the last few years, IBM has invested massive resource in IBM Watson Health, its computing-based healthcare division. IBM hopes Watson Health will play a key role in healthcare’s coming “cognitive era,” a time of profound industry transformation.

Rometty is far from alone in her enthusiasm, as investors are throwing massive amounts of capital at healthcare AI startups. According to CB Insights, more than 100 healthcare companies are using AI to offer virtual assistance to patients, shorten drug discovery times, crunch clinical data, and more.

The explosion of healthcare AI startups is just beginning, with predictions that AI and cognitive computing will generate $6.7 billion in revenue by 2021. And that’s not all—according to Frost & Sullivan analyst Harpreet Singh Buttar, “by 2025, AI systems could be involved in everything from population health management, to digital avatars capable of answering specific patient queries.” But the vision for AI extends beyond a local level, with Buttar noting that, “On a global scale, in regions with high underserved patient populations, AI is expected to play a significant role in democratization of information and mitigating resource burdens.”

On the surface, it seems all but inevitable that AI will become healthcare’s “nervous system,” as giant consulting firm Accenture recently declared. On the other hand, if you’ve watched technology trends long enough, you know that grand predictions don’t always pan out and that breathless excitement often deflates over time. As with blockchain, another emerging technology with a sexy reputation in health IT circles, it’s hard to tell where healthcare AI will be when the frenzy dies down.

A history of hype

Artificial intelligence. Machine learning. Cognitive computing. Given the fevered pitch of interest in these technologies, you’d think they were invented yesterday. The truth is, though, that they’re far from new—AI development goes back decades, and over that time, it’s gone in and out of fashion. “The history of AI is one of long ‘winters’ of disinterest punctuated by brief periods of hype and investment,” one editor notes.

The birth of artificial intelligence is said to have taken place in the mid-1950s, when a workshop at Dartmouth College sparked the birth of AI research. At the time, attendees predicted that a machine as intelligent as a human being would exist in more than a generation. While they were given millions of dollars to bring this to fruition, the researchers hit a brick wall, as they hadn’t taken the limitations of the era’s computing hardware into account. By 1973, the U.S. government had stopped funding broad AI projects.

Since then, interest in AI has gone through a boom-and-bust cycle, as researchers and technology innovators have worked to solve critical problems with AI technology. As with the hopefuls at the fateful Dartmouth meet-up, each has pushed their generation of computer hardware and software as far as they could, but typically hit a limit over time.

Now, we’re clearly back in euphoria mode. In healthcare, important stakeholders like hospitals, doctors, and pharmaceutical companies believe that AI will change the way their industry works. Venture capital firms are raining money down on healthcare AI startups. While we may or may not be looking at a hype cycle—that’s a matter of interpretation—we’re clearly experiencing something that’s different from business as usual.

AI’s success stories

While the industry is in its early days, there’s already a trickle of healthcare AI success stories emerging. As you might guess, many promising technologies aren’t getting as much attention as Watson, but smaller companies are making strides.

Take Babylon Health, a AI-based mobile health company backed by Google’s DeepMind Technologies. Early results from a pilot using its AI-based chatbot to triage patients with the UK’s National Health Service suggest that it may be as good as clinicians at sorting out patients. If healthcare pros agree on any single approach to AI use, it’s that anything that frees doctors from mundane, repetitive work to do more advanced tasks is good for everyone.

In a more tech-driven use case, Medicrea, the first company in the United States to gain a FDA clearing for personalized spine products, uses powerful algorithms to measure individual spinal configurations. Using that information, a personalized titanium or cobalt chromium rod is created to restore that person’s spine to it’s optimal shape. This unparalleled  and AI-driven precision allows Medicrea to offer a lifetime guarantee on their technology and procedures.

As covered in our latest Last Week in Health Tech, even small companies are finding ways to utilize AI in meaningful and effective ways—Wysa, an AI-driven chatbot, uses cognitive behavioral therapy techniques to engage with people suffering from mental illness. With over one million active users, Wysa is proving that applications of AI don’t need to be grandiose to be effective.

The list of use cases for healthcare AI tools is expanding by the day, and machine learning algorithms may soon help providers predict which patients might be at risk for diseases, not following care plans, or experiencing frequent hospitalizations. Not only that, when connected with EMRs via an API, virtual clinicians will be able to scan a patient’s entire health history and offer patients well-informed recommendations on managing their health.

Leaving a mark

With the list of well-funded startups growing, and the interest of technology giants increasing, it seems a given that we can expect more healthcare AI wins. This is not to say that the level of excitement around healthcare AI won’t peter out over time. In fact, if history is any guide, some trend chasers will undoubtedly lose interest over and move on to the next most-promising technology.

But even if healthcare AI failures are rife, the real visionaries will lead and the tech creators will continue to invent. After all, even the spectacular boom and bust of the dotcom era left technologies in its wake which shaped the Internet we know today. Hype or no, healthcare AI will leave its mark.