Branded as a joke by venture capitalist Chamath Palihapitiya, founder and CEO of Social Capital, IBM’s big bet on enterprise AI has garnered more bad PR than any other AI system. Today, it is under fire from NYC headquartered institutional securities firm Jeffries’s for being too expensive and an underperformer compared to other leading big shots such as Amazon, Microsoft, Facebook, Alphabet and Apple.
One of the most ascribed words to IBM’s core growth unit -Watson- is overhyped and while it may have succeeded in powering specific business use cases of Fortune 500 companies, there is one that burned it particularly – Texas based MD Anderson Cancer Centre that shelved the project after reportedly spending $60 million on the technology.
Watson has a huge list of detractors. Its biggest criticism is coming in the form of its inability to justify the investment to shareholders, with some succinctly terming it as the “Cadillac of the AI world”. In other words, the system requires a lot of data preparation. One of the key problems from the MD Anderson case was that the Cancer centre had changed software systems and the AI system was unable to interact with the new database without requiring any additional integration work.
Making a case for Watson
However, before we drub Watson outright, here’s our counterview. According to a latest study, the results from Manipal Comprehensive Cancer Centre in Bangalore, India have been positive and hold much promise in paving the way for evidence-based treatment decisions. The findings indicate that treatment recommendations made by the Manipal multidisciplinary tumour board and Watson for Oncology were highly concordant in the cancers examined. This data was presented at the American Society of Clinical Oncology meeting, the first data presented at a medical conference and also included findings from Memorial Sloan Kettering.
According to Shalini Kapoor, Engineer and Chief Architect, IBM Watson, India is an important market with domains such as healthcare, finance, telecom, education among others are ripe for Watson. It’s true Watson may suffer from overzealous market, however, it’s shortcomings are largely attributed to the hyperbolic claims rather than the technology underpinning of the cognitive system. In the MD Anderson Cancer Centre case, Watson’s shortcomings were not hinged on its technical or functional capabilities, but the inability to assimilate in hospital environment. According to excerpts from Journals of the National Cancer Institute, an expert described the application’s problem in integrating case reports and other text-heavy medical information. Watson is also working with several Indian startups such as Textient Analytics, Talview and Trans Neuron Technologies among others in delivering insights.
The verdict according to American Consulting firm Booz Allen Hamilton firm is that most machine learning systems have yet to become mature and sophisticated to deliver clinical decisions. Another big area of concern is that in healthcare, there are a lot of privacy concerns and analysts believe it remains at best, a research project.
Counterview – why Watson should be valued
IBM CEO Ginni Rometty once famously said at a World of Watson 2016 conference that in another five years, IBM Watson will be behind every major decision – personal or business. This may seem like an overstatement today, but Watson’s contribution to the cognitive and AI world is immense.
- Firstly, Watson is credited for making AI the new normal, carrying the technological capabilities of the cognitive system into the general business. In other words, IBM turned Watson into a tool for business, providing an Enterprise AI solution in the market ripe for AI. Watson’s cognitive capabilities caught fire, thanks to the marketing sell that created a cloud of hype around it.
- While other behemoths Facebook Amazon and Google are bent on making AI work for their internal divisions, IBM opened the Jeopardy winning supercomputer for general business. Although, Google has made significant headway into healthcare with research related to diabetic retinopathy, Facebook and Apple remain secretive about their AI systems. Google is also imbuing its own platforms and products with AI, with the company pivoting from a mobile-first to AI-first strategy.
- IBM Watson clearly made the first foray in General AI and has plenty of successful implementations, including healthcare, writes Vijay Vijayasankar, General Manager, Cognitive/IOT/Analytics/Watson Health, IBM GBS North America – IBM.
- IBM Watson is no longer the only game in town. You don’t need a crystal ball to predict that. Watson is delivered as a set of APIs that the ecosystem can use easily. Today, IBM clients are building cognitive applications via Watson. And Vijayasankar believes soon somebody or IBM will package it. And Just like Watson, its competitors (Amazon, Microsoft, Google et al) also have a lot to prove in the enterprise AI market. At least, Watson’ moonshot for healthcare still holds promise as Watson has been trained on 8 types of cancer and is deployed in six countries.
- Thanks to IBM’s over-the-top focused articulated marketing strategy, billions of dollars of investment were channeled into the field, thereby helping the ecosystem thrive. IBM clearly shunted AI from the R&D labs into the business world and today in healthcare, Watson is used by the half of the top 25 life science companies.
Was IBM Watson mis-sold on the marketing front?
Today, IBM Watson is challenged on all fronts. It’s facing intense competition from big and small players in the AI space, on the costing front and acute talent shortage in artificial intelligence field. When it first launched, IBM Watson rode high on the AI wave wooing CTOs with a lot of feverish excitement around its cognitive capabilities that spanned every domain. Some of Watson’s much-feted abilities are digesting unstructured data to produce insights, domain-related understanding of business problems and forming the AI backbone of an enterprises. While it is still too early for customers to determine the results, one thing is abundantly clear, Watson always had a targeted, focused marketing push and this may have led not just to overselling, but mis-selling as well.
Here’s what we think: today Artificial Intelligence systems deployed can be utilized effectively to pare down routine and time-consuming jobs, thus allowing doctors and medical staff to focus on more intensive work. Just the same way Google’s retinal screening AI system that debuted in Aravind Eye Hospital is doing screening work, freeing up ophthalmologists to see more patients. The commercial AI market is no longer Watson’s stomping ground, big players like Amazon, Apple and Microsoft are also chipping away at its heels with better solutions. That is why we are seeing the likes of Microsoft and not just Google peering into the healthcare space. On the home-front as well, Watson has plenty of competition from the likes of Infosys Mana & Wipro Holmes.
Also, IBM’s recent tie-up with Automation Anywhere, leading developer of Robotics Process Automation software marks the company’s foray in intelligent automation. In other words, it means that IBM is planning to expand the solutions portfolio with RPA capabilities. But where does this leave Watson? Seems like IBM needs to get its digital strategy together.
Try deep learning using MATLAB