IBM Doubles Down on AI
Despite the attraction of consumer-facing AI, IBM's focus on practical business applications may yet win the race for widespread AI adoption.
- By Barry Devlin
- January 18, 2017
In a previous article, I looked at developments from Google and its subsidiaries in the field of artificial intelligence (AI). The work is notable for its worthwhile advances in the underlying technology and algorithms of AI and machine learning and their application to specific, personally meaningful human tasks, such as map navigation and language translation.
In an insightful, must-read article about Google's journey into artificial intelligence --"The Great A.I. Awakening" in the New York Times Magazine -- Gideon Lewis-Kraus spins three threads of AI evolution. "The seven-decade story is about what we might conceivably expect or want from it. The five-year story is about what it might do in the near future. The nine-month story is about what it can do right this minute. These three stories are themselves just proof of concept. All of this is only the beginning."
Google's approach, and that of other millennial companies at the forefront of AI, such as Facebook, Apple, and Amazon, is to directly address the ordinary person who, smartphone in hand, needs to get somewhere quickly, find that special gift, or just be instantly entertained. From virtual reality headsets to autonomous vehicles, the glitter of AI is sprinkled everywhere, and the public is enthralled. Star Trek magic is just around the next corner.
There is another side to AI that is less consumer-oriented and more "serious." In contrast to Google and its ilk, IBM's AI story may appear more mundane, less expansive. Once a behemoth of the IT industry itself, IBM's research pockets are now shallower than Google's and its goals correspondingly narrower.
IBM targets opportunities and problems that impinge more directly on business goals and financially valuable decisions. At times this extends even to the unsexy field of data management.
Building on the IBM Watson initiative and success, IBM uses the term cognitive computing for its market messaging. The difference between AI and cognitive computing seems nuanced at best, so let's roughly equate them for now. Indeed, IBM has also rebranded AI as "augmented intelligence" -- as explained in an IBM Research response to a 2016 request for information "Preparing for the Future of Artificial Intelligence" from the White House Office of Science and Technology Policy.
With this shift of emphasis, IBM aims to "enhance and scale human expertise" rather than replace humans with automated systems. Metaphorically, the focus moves from artificial intelligence replacing the driver in an autonomous vehicle to augmented intelligence benefitting all travelers in the transportation infrastructure.
Examples of Augmented Applications
In business terms, IBM's AI applications have thus been "augmented" in nature. In healthcare, Watson's earliest and probably best-known work, deep knowledge is gleaned from the expertise of top consultants, analysis of the entire corpus of medical literature, and access to detailed (anonymized) symptomatic and genomic patient data. This allows expert, personalized diagnosis and treatment recommendations to be offered to more patients in more locations via local, less specialized doctors.
IBM works in close partnership with leading institutions and companies such as the Mayo Clinic and Memorial Sloan Kettering (MSK) Cancer Center to embed the best existing knowledge and to roll out the results to the wider community. As Mark Kris, a medical oncologist at MSK, states: "If it doesn't get to people who benefit, it's just irrelevant."
As opposed to the theoretical AI push and personal application approach of Google, IBM is leading the way in practical application at the enterprise and societal level.
In the legal world, Watson powers the world's first "artificially intelligent attorney" that can search and analyze thousands of pages of legislation and case law in minutes. It can evaluate how opponents argued other cases to offer opinions on what the law is and how it is likely to be interpreted in particular areas. It can also improve its performance by "learning" from feedback. Once again, partnerships with law firms and providers of legal technology are key to the IBM approach.
Moving to Decision Support
Applying IBM Watson technology to business analytics and decision-making support is a work in progress, but it promises to transform an increasingly complicated environment where data scientists currently work in disconnected silos with diverse tools.
IBM Watson Data Platform applies AI/cognitive computing to the tasks of finding, integrating, and governing data, as well as making it available for task-specific, collaborative use across the organization. With the focus firmly on a platform, IBM is aiming to make big data and machine learning widely useful and usable rather than developing new AI technology or algorithms.
Although the attraction of AI-driven autonomous vehicles on the streets and all-knowing assistants in the home is undeniable, providing better support for decision making at all levels of business is a long-standing and thus far poorly met requirement in every industry. IBM's tortoise to Google's hare may yet win the race for widespread and practical AI adoption.
Dr. Barry Devlin defined the first data warehouse architecture in 1985 and is among the world’s foremost authorities on BI, big data, and beyond. His 2013 book, Business unIntelligence, offers a new architecture for modern information use and management.