Most people know Watson as (1)IBM’s answer to Jeopardy star Ken Jennings. But IBM’s (2)ambitions for its (3)artificially intelligent supercomputer are now less quiz show champion and more medical genius.
Watson, the supercomputer that is now the world Jeopardy champion, basically went to medical school after it won Jeopardy. MIT’s Andrew McAfee, (4)coauthor of The Second Machine Age, said recently in an interview with Small Planet, “I’m (5)convinced that if it’s not already the world’s best (6)diagnostician, it will be soon.”
Watson is already capable of (7)storing far more medical information than doctors, and unlike humans, its decisions are all (8)evidence-based and (9)free of (10)cognitive biases and (11)overconfidence. It’s also capable of understanding (12)natural language, (13)generating (14)hypotheses, (15)evaluating the strength of those hypotheses, and learning ー not just storing data, but finding meaning in it.
As IBM scientists continue to (16)train Watson to (17)apply its vast stores of knowledge to actual medical decision-making, it’s likely just a matter of time before its (18)diagnostic (19)performance (20)surpasses that of even the sharpest doctors.
Back in 2011, McAfee wrote on his blog about why a diagnosis from “Dr.Watson” would be a (21)game changer:
(A) It covers all (22)available medical knowledge. Human doctors (23)can’t possibly hold this much information in their heads, or (24)keep up with it as it changes over time. Dr.Watson knows it all and never (25)overlooks or forgets anything.
(B) It’s accurate. If Dr.Watson is as good at medical questions as the current Watson is at game show questions, it will be an excellent diagnostician indeed.
(C) It’s (26)consistent. Given the same inputs, Dr.Watson will always output the same diagnosis. (27)Inconsistency is a surprisingly large and common mistake among human medical professionals, even (28)experienced ones. And Dr.Watson is always available and never annoyed, sick, nervous, (29)hungover, (30)upset, (31)sleep-deprived, or so on.
(D) It has very low (32)running costs. It’ll be very expensive to build and train Dr.Watson, but once it’s up and running, the cost of doing one more diagnosis with it is (33)essentially zero, (34)unless it orders test.
(E) It can (35)be offered anywhere in the world. If a person (36)has access to a computer or mobile phone, Dr.Watson is (37)on call for them.
Watson has read dozens of textbooks, all of PubMed and Medline (two (38)massive databases of medical journals), and thousands of patient records from Memorial Sloan Kettering Cancer Center. (39)Altogether, “Watson has analyzed 605,000 pieces of medical evidence, 2 million pages of text, 25,000 training cases and had the assistance of 14,700 (40)clinician hours (41)fine-tuning its decision accuracy,” Forbes reported in 2013.
And it’s getting “smarter” every year. So how would Dr.Watson work (42)in practice? Here’s how IBM describes the process:
First, the (43)physician might describe (44)symptoms and other (45)related factors to the system. Watson can then (46)identify the key pieces of information and analyze the patient’s data to find (47)relevant facts about family medical history, current (48)medications and other existing conditions. It combines this information with (49)current findings from tests, and then forms and tests hypotheses by examining a variety of data sources. From here, Watson can provide potential (50)treatment options.
The supercomputer’s potential is huge, but ー as The Wall Street Journal reported earlier this year ― currently “just (51)a handful of customers are using Watson in their daily business,” and it’s far from performing at the level and in the range of domains that should be possible in the future.
(52)So far, IBM’s most (53)high-profile AI (Artificial Intelligence) partnerships are with MD Anderson Cancer Center, where Watson helps recommend (54)leukemia treatments, and WellPoint, where Watson helps the (55)insurer evaluate doctors’ treatment plans. WellPoint claims that the system is already significantly better than human doctors at diagnosing (56)lung cancer.
Watson is not yet able to use all the information it has absorbed, so it still has a long way to go before it (57)catches up with our best human diagnosticians, whose (58)versatility and (59)agility are difficult to (60)match. But Watson’s ability to learn, analyze, and apply knowledge suggests that it will (61)get there ― (62)eventually.
(Bisiness Insider, April 22, 2014, modified)