SEOUL -- Samsung Electronics is attempting a rapid buildup of its artificial intelligence program led by chief Lee Jae-yong to catch up with front-runners like Google and diversify an earnings base skewed toward chipmaking.
By 2020, the South Korean manufacturer aims to have 1,000 AI developers working in five countries.
In May, Samsung appointed ex-Google AI developer David Eun as its first chief innovation officer, according to a source familiar with the matter. Since joining the South Korean company about six years ago, Eun has forged partnerships with startups and worked on investments.
Samsung did not officially announce the creation of the new post, but an executive said the CIO's role is to "cultivate new businesses of the future on themes like AI and the 'fourth industrial revolution,'" using a term for networked machines.
That same month, Samsung opened AI research centers in the U.K., Canada and Russia, joining an existing hub at home and one in the U.S. Each center will capitalize on research strengths and connections in its country, with the British hub to work on data analysis, the Canadian one voice recognition, and the Russian one physics. Samsung does not reveal the current size of its AI engineering staff but plans to have 600 technicians in South Korea and 400 abroad by 2020.
In March, soon after being released from prison time served on charges of corruption involving South Korea's ousted ex-President Park Geun-hye, Lee traveled to Europe, the U.S. and the rest of Asia to connect with executives and experts in the AI field. The de facto head of the Samsung conglomerate also is believed to have visited Japan and Hong Kong this month, though he has kept a low public profile.
Also in June, Samsung said two prominent American AI researchers, Sebastian Seung of Princeton University and the University of Pennsylvania's Daniel Lee, would leave their academic posts to join its ranks.
Samsung offers AI features in its devices, but the reportedly spotty performance of the voice recognition function of its high-end Galaxy S9 smartphone, released in March, suggests the company still lags Google in this regard.
Even with the buildup, Samsung will not measure up to leading U.S. tech companies, whose AI research forces number in the thousands, but it will pull even with big multinational manufacturers making their own way in the technology. Toyota Motor, for instance, said in March it was forming an AI unit focused on automated driving with a team of around 1,000, led by former Google robotics engineer James Kuffner.
With so many players competing for talent, recruiting staff will pose a challenge for Samsung.
In fields like liquid-crystal displays and semiconductors, Samsung rapidly gain technological prowess and market share, in part by luring staff from Japanese companies with generous terms. Having booked an operating profit of about 53 trillion won ($48.8 billion) for 2017, it may reach into its deep pockets once more to attract talented AI technicians.
Samsung aims to develop so-called AI chips that can handle data three to four times faster than conventional chips, to enable core AI functions like deep learning. Such chips are expected to play a key role in making ultra-fast 5G communications practical, as well as the technological advances that accompany these new networks, by speeding up processing of videos and other data.
The semiconductor segment provides more than 70% of group operating profit at Samsung, which unseated U.S. chipmaker Intel as the global revenue leader in the industry last year. Earlier this year, the South Korean company also announced big investments in contract chipmaking. As Apple and other tech companies increasingly design their own chips suited to their particular needs, Samsung hopes AI advances will help expand its contract chipmaking business to take on global leader Taiwan Semiconductor Manufacturing Co.
Samsung also aims to apply AI to making autoparts for self-driving cars, as well as helping discover new drugs. The hope is that the techology will inject vigor into two businesses that have shown slow growth despite efforts to turn them into major income streams.
DRAM, which generates 90% of the semiconductor segment's profit, has become the target of Chinese antitrust probes, and the South Korean company was raided on May 31, according to a source familiar with the matter. Given this and a generally hazy outlook for the DRAM market, the company is racing to cut reliance on memory chips.