Page 61 - The Economist
P. 61
Business The Economist December 9th 2017 61
Also in this section
64 Health care in America
64 Gaming and gambling
65 Foreign brand names in China
66 Rio Tinto, robominer
67 Schumpeter: Walmart fights back
For daily coverage of business, visit
Economist.com/business-finance
Artificial intelligence Fuelled byrivalry, high hopesand hype,
Battle of the brains the AI boom can feel like the first Califor-
nia gold rush. Although Chinese firms
such as Baidu and Alibaba are also invest-
ing in AI, and deploying it in their home
market, the most visible prospectors are
Western tech firms. Alphabetiswidely per-
ceived to be in the lead. It has been making
SAN FRANCISCO sizeable profits from AI for years and has
Tech giants are investing billions in a transformative technology
many ofthe best-known researchers. But it
OMMANDING the plot lines of Holly- nique isnowused in all mannerofapplica- is early days and the race is far from over.
Cwood films, covers of magazines and tions in the tech industry, including online Overthe nextseveral years, large tech firms
reams of newsprint, the contest between ad targeting, product recommendations, are going to go head-to-head in three ways.
artificial intelligence (AI) and mankind augmented reality and self-driving cars. They will continue to compete for talent to
draws much attention. Doomsayers warn Zoubin Ghahramani, who leads AI re- help train their corporate “brains”; they
that AI could eradicate jobs, break laws search at Uber, believes that AI will be as will try to apply machine learning to their
and start wars. But such predictions con- transformative as the rise ofcomputers. existing businesses more effectively than
cern the distant future. The competition to- One way to understand AI’s potential rivals; and theywill tryto create new profit
day is not between humans and machines impact is to look at databases. From the centres with the help ofAI.
but among the world’s technology giants, 1980sthese made itcheap to store informa-
which are investing feverishly to get a lead tion, pull out insights and handle cognitive Idiot savants
overeach otherin AI. tasks such as inventory management. The most frenzied rush is forhuman talent,
An exponential increase in the avail- Databases powered the first generation of which is farmore scarce than eitherdata or
ability of digital data, the force of comput- software; AI will make the next far more computing power. Demand for AI “build-
ing power and the brilliance of algorithms predictive and responsive, says Frank ers” who can apply machine-learning
has fuelled excitement about this formerly Chen of Andreessen Horowitz, a venture- techniques to huge sets of data in creative
obscure corner of computer science. The capital firm. An application such as Goo- ways has ballooned, far exceeding the
West’s largest tech firms, including Alpha- gle’s Gmail, which scans the content of e- number of top students who have studied
bet (Google’s parent), Amazon, Apple, Fa- mails and suggests quick, one-touch re- the techniques.
cebook, IBM and Microsoft are investing plies on mobile devices, is an early exam- Today AI systems are like “idiot sa-
huge sums to develop their AI capabilities, ple ofwhat could be coming. vants,” says Gurdeep Singh Pall of Micro-
as are their counterparts in China. Al- As with past waves of new technology, soft. “They are great at what they do, but if
though it is difficult to separate tech firms’ such as the rise ofpersonal computers and you don’t use them correctly, it’s a disas-
investments in AI from other kinds, so far mobile telephony, AI has the potential to ter.” Hiring the right people can be critical
in 2017 (see chart1on nextpage) companies shake up the businesses of the tech giants to a firm’s survival (some startups fail for
globally have completed around $21.3bn in by helping them overhaul existing opera- lackofthe rightAIskills) which hasset off a
mergers and acquisitions related to AI, ac- tions and dream up new enterprises. But it trend of firms plundering academic de-
cording to PitchBook, a data provider, or also comes with a sense of menace. “If partments to hire professors and graduate
around 26 times more than in 2015. you’re a tech company and you’re not students before they finish theirdegrees.
Machine learning is the branch of AI building AI as a core competence, then Job fairs now resemble frantic “Thanks-
that is most relevant to these firms. Com- you’re setting yourself up for an invention giving Black Friday sales at Walmart”, says
puters sift through data to recognise pat- from the outside,” says Jeff Wilke, chief ex- Andrew Moore, dean of Carnegie Mellon
terns and make predictions without being ecutive of“worldwide consumer” at Ama- University’s(CMU) school ofcomputersci-
explicitly programmed to do so. The tech- zon, and adjutant to JeffBezos. ence, a pioneering institution in AI (whose 1