12.01.16

When Watson Doesn’t Compute: 3 Myths Debunked

IBM’s Watson is using artificial intelligence to “learn” healthcare. 

While it holds undeniable potential to improve patient care, Watson’s ability to deliver on HIT’s ultimate goal of liquid data and interoperability is unclear. Signet Accel CEO, John Raden shares his point of view:

While Watson’s future is laden with possibility, the question is, is Watson a good fit for your work? Do you need Watson to solve your challenges, or do you need interoperability among the hard-won data of your own organization and its collaborators? It’s important to evaluate your own goals and challenges and remain mindful that Watson’s knowledge isn’t inherent. In order to provide true and lasting value in healthcare, it needs your data. But what does your organization need?

Raden’s viewpoint debunks Watson-related myths in three key areas:

Myth: With Watson, your organization will maintain ownership of your data. While some HIT solutions allow organizations to stay in control of their hard-won patient data, Watson’s AI doesn’t fit the mold. Data integration is not its sole purpose, and your data ownership is not its concern. It requires users to relinquish it so the information can be centralized and assimilated into Watson’s intelligence. The catch? Your data becomes property of Watson and available to all of its users as well.

Myth: Centralizing your data is the best way to integrate your data. This is the great untruth in our field that has kept many of us mired in the costly, time-intensive work of relocating disparate data into one location in order to integrate it. In reality, centralization is a barrier to progress, both in sharing and security. Unfortunately, the drawbacks don’t end there. Centralization can also result in data duplication, removal of adequate permissions and loss of data fidelity—ultimately strangling data liquidity.

Myth: Watson provides interoperability for seamless data integration. In truth, Watson provides access to broad-spectrum information. You ask Watson a question, and it searches its intelligence and makes its best guesses at the answer. It’s also an excellent data analytics tool. However, it does not fulfill the promise of true interoperability—integrated raw data viewable and accessible across organizations, regardless of its language or origin. Data that’s easy to access and simple to query. If interoperability is your goal, Watson is not the way.

Herein lies the question: Is Watson the right investment for your organization? 

Find answers and more in the John Raden resource, When Watson Doesn’t Compute.

New Call-to-action