Imagine this: your symptoms paint a dire picture and your GP is worried; it could be a cancer. “Let’s talk to an oncologist,” she says.
In New Zealand, like most countries, you’ll probably join a queue to see the specialist. Could be days, possibly weeks. Meanwhile, that tumour keeps on growing.
But instead of waiting anxiously, now imagine your GP turns to the computer on her desk and starts to talk to Watson, IBM’s much-vaunted artificial intelligence machine. The doctor has already uploaded your medical history, x-ray, CT-scan, bloods, dietary habits, family history and even your scribbled diary notes; so Watson is well prepared.
“Tell me Watson, what’s going on here?” she asks.
Comparing your own condition with millions of other patients and 25 million medical journal articles, Watson comes back within moments. “Based on what we’re seeing, your patient has a 15% chance of a tumour and 90% likelihood of an irritable bowel. I’d recommend a blood test seeking a t-cell count.”
“Well that’s better than we thought!” says the GP. And you’re off to the lab for a jab in the old jacksie.
That scene could come straight from the PR books of IBM because that’s the future Big Blue hopes will happen. It’s not without merit.
In the last two years Watson has digested millions of articles from medical journals, thousands of patient records and has been trained by doctors at the highly-regarded hospitals such as Memorial Sloan Kettering (MSK) and Anderson Cancer Center in the USA. As a result, Watson could be the most knowledgeable oncologist in the world.
I travelled to New York as a guest of IBM to meet Watson, or at least the team behind it and it’s fair to say the global giant is excited about its new toy. Very excited.
So, who or what is Watson Health?
It’s not a doctor. And it’s not a he. Let’s get that straight. Watson Health is an overarching brand for a suite of cloud computing services aimed at the health sector. The brand leverages the computing power in the Watson system for:
- Natural language input and output
- Storage and processing of unstructured data (like doctors’ notes)
- Millions of data points; images, journal articles, patient records, Fitbit records
- Massive data processing power
- Global network of sales and collaborators
- And an open network for developers
Here’s a video of it:
And another one that’s a bit less PR-ish:
So has IBM gone all medical now? No. It’s promoting Watson as just one part of a ‘Health Cloud’ where it provides services to health companies, researchers and hospitals. You can probably imagine at this point that they’ll use the world ‘eco-system’ and sure enough they have a ‘partner eco-system’ including the afore-mentioned MSK, the Bunrungrad Hospital in Thailand and companies such as:
- CafeWell – developing an app for analysis of menus for the diet-conscious
- TalkSpace – an app for ‘global constant care’ and education
- Interactive – long term care planning between patient and hospital
- ThinkGenetic – Q&A in plain English about genetic disorders
- GenieMD – mobile patient data, held by patients but available to physicians
It’s also done work with Under Armour to prototype a Fitbit style device and Medtronics to create a Bluetooth-connected insulin pump.
It sounds an expensive gamble. It is. The company claims to be investing US$1 billion in Watson, including $100 million on the fancy offices I visited in Astor Place, New York. (Yes, they were cool for IBM but not as cool as these).
A lot is riding on Watson. I met with Ashish Cowgali, one of IBMs super brains who said that Watson really is a ‘moonshot’ for the company. Most commenters reckon IBM is struggling to make the shift from providing corporate consulting services to the software-as-a-service world of cloud computing. In this area it’s up against a bunch of global giants such as Google and Amazon and also nimble specialists like our very own Orion Health.
To make the most of its new AI, IBM wants to deploy Watson as a SAAS function into medicine, science, insurance, retail, heck, any industry that its customers are in.
Is it working?
It’s too early to tell if Watson Health is delivering on its promises. A recent Financial Times article contained scathing criticisms from industry analysts who say that Watson in general, and Watson Health in particular, has yet to deliver tangible outcomes for clients. Oncology remains troublesome.
Says the FT:
“The University of Texas’s MD Anderson Cancer Center began trying to train the system three years ago to discern patients’ symptoms so that doctors could make better diagnoses and plan treatments.
“It’s not where I thought it would go. We’re nowhere near the end,” says Lynda Chin, head of innovation at the University of Texas’ medical system. “This is very, very difficult.” Turning a word game-playing computer into an expert on oncology overnight is as unlikely as it sounds, she says.”
Here’s a summary of the FT piece:
What does it all mean?
For close observers of IBM and the health sector, the experiment in Watson looks bold and risky and not immediately disruptive. It also suggests that what IBM is calling ‘cogntive’ is probably just pretty smart and fast computing, not truly AI. That particular beast is yet to be truly unleashed.
For the rest of us who wonder if cancer will ever be cured, IBM will be a significant player and machines like Watson will be at a clinic soon. Maybe not today, but pretty soon, the computer will see you.
Vincent Heeringa is Idealog’s publisher-at-large. He travelled to New York courtesy of IBM.