Hitachi: practical AI or humanoid robots?
Posted: 7 December 2017 | By Charlie Moloney
What are Hitachi doing with artificial intelligence (AI) technology? Are they building practical AI for enterprise or are they building humanoid robots?
“Transport, energy, public safety through video analysis, and operational efficiencies”, Ram Ramachander, the CCO of the Social Innovation Business at Hitachi Europe, told Access AI, during the Hitachi Social Innovation Forum, held in Knightsbrige, London last Thursday.
Mr. Ramachander said, “Hitachi’s involvement in digitization, AI, comes from it’s digital engineering pedigree, and we digitize ourselves first before we take it to our customers.”
Hitachi are digitizing their rail network by installing sensors on all their trains which will collect data on what is happening inside and outside as they move at high speeds, and then using AI to gain insights which could lead to greater operational and energy efficiency.
I think even we get confused sometimes, we talk like IBM when we’re not. What we are is, we’re pragmatic users of AI
“We’re not necessarily going to advise our customers to do what we’re doing”, Mr. Ramachander said, “but we’re going to take all of the knowledge that we gain from digitizing our rail infrastructure into the automotive sector as a whole.”
Looking at the rest of the iceberg
Hitachi have been helping customers go digital by combining AI, the Internet of Things (IOT), and Big Data to optimize their operations.
They claim to have saved the largest, public facing US transportation logistics company (name not disclosed) $18.2M a year, and South West Water, an English company providing drinking water and waste water services, reduced customer leakage cases by 24%.
However, if the number of results that come up on Google is any indicator of anything, then Hitachi is not the company that we mention when we talk about AI.
Companies like Facebook, Oracle, Alibaba, and IBM dominate the headlines and the public imagination with stories of machines that are beating humans at their own game and breaking free of our control.
“We’re not IBM”, Mr. Ramachander said, “and I think even we get confused sometimes, we talk like IBM when we’re not. What we are is, we’re pragmatic users of AI. We can tell you where it can really deliver value.”
“The difference between a pure play technology company and a digital engineering company like us is that it’s very hard for the technology company to get the domain expertise to really understand what people mean when they say they want to optimize their manufacturing line.”
“IOT, AI, and Big Data is like an iceberg, the technology ennoblement is just the tip of the iceberg, the bit underneath is the deep domain capability and knowledge that we’ll be bringing to the table to make that transformation really happen.”
The little red and white robot: EMIEW3
One thing not mentioned in the Hitachi mission statement that Mr. Ramachander outlined is the cute, humanoid robot which is the face of the Social Innovation team’s ad campaign.
In the ad, the robot rushes to the aid of a lost traveller and ferries him to a waiting cab, ensuring that the man arrives just in time for his daughter’s piano recital.
This robot is called EMIEW3 and was released in April 2016. It is superior to its predecessor, EMIEW2, because it is connected to an IoT network, giving it the ability to access information, such as directions.
Dong Li, a Research Engineer at Hitachi Europe, said, “We are working on making it recognise people, and potentially the emotions of people. We envisage EMIEW3 could potentially work in transport hubs where it could identify passengers that seem to be lost and could proactively approach the passengers and offer help.”
EMIEW3 was tested in Hanada airport in Japan in 2016, but whether the robot will ultimately work at transport hubs is still undecided.
All robots which use natural language processing (NLP) for customer service face the following issues: siphoning off background noise, understanding casual speech and different accents, and pleasing customers which may not be comfortable in being served by a machine.
One can imagine a version of the Hitachi advert showing an increasingly agitated traveller dealing with a robot which couldn’t understand him until he finally missed his flight.
Leonard Ah Kun, a Hitachi Europe UX designer, said that Hitachi are looking at deploying EMIEW3 in shopping malls, where there is less urgency than in airports.
“We do have customers that have approached us”, Mr. Ah Kun said, “They understand that the hardware and software is slowly getting better. They’re open to starting with solutions that are simple.”
“It’s not necessarily about the robot on its own. If it worked in a retail environment alongside our image processing technology then EMIEW3 could know where you have been, and might be better able to help you. That’s why EMIEW3 would need to operate as a part of the IOT environment.”
Dong Li, a Research Engineer at Hitachi Europe, showcased EMIEW3 to Access AI at the Hitachi Office in London
The ROI in the machine
Hitachi do not yet know for sure whether EMIEW3 in its current form will be a commercially viable product.
They’re currently working on a Digital EMIEW project where they can give potential customers demos of the robot in test scenarios quickly, and without the additional overhead of showing the hardware.
Using augmented reality vision, customers can walk around their office and EMIEW3 can navigate them.
The obvious question is: why even have the robotic body, if you could project a virtual EMIEW3? The answer is not yet clear.
There is no robotics company which isn’t struggling to prove the value of consumer-facing bots like EMIEW3. For now, companies with humanoid robots are making their return on investment (ROI) in bombastic publicity stunts.
The Sophia robot, built by Hanson robotics, was granted Saudi citzenship in October, and Pepper, the Softbank robot, has had chanting software written for it so it can perform Buddhist funeral ceremonies.
When asked whether Hitachi wanted to thrust EMIEW3 into the limelight in a similar way, Mr. Ramachander said, “No, we’re pragmatic”.
“We in Social Innovation will identify EMIEW3 as part of the value chain, connect it to the ecosystem, and then we’ll launch it. But it’ll be a part of an entire delivery package, rather than on its own. We’re not trying to prove that we’re the best Android developers”.