kwanchaift - stock.adobe.com
IoT and artificial intelligence dominate APAC tech landscape in 2017
Businesses across the APAC region have started to dip their toes into AI and the internet of things to improve operations and stay competitive
In their quest to get smarter in the way they operate in the face of tightening IT budgets and a lack of skilled manpower, companies in the Asia-Pacific (APAC) region have started to dip their toes into the internet of things (IoT) and artificial intelligence (AI).
Take Singapore’s leather goods retailer Goodvine Group, for example. At its warehouse, smartphone-toting workers can complete a stock-take of a quarter of all items in less than half an hour.
Over in Japan, Keifuku Bus, which operates a fleet of 200 public buses serving the Fukui prefecture, has equipped its drivers with sensor-laden shirts to keep a lookout for those with signs of fatigue in a technology trial.
Such technology-infused workplaces, dotted with IoT devices, are fast becoming the norm across the APAC region, according to research by Zebra Technologies, with seven out of 10 APAC retailers planning to invest in IoT technologies by 2021 to address customers’ bugbears, such as the inability to find an item.
Although IoT is slowly gaining ground across the region, the adoption of the technology is still in its infancy, at least in Southeast Asia.
In a survey conducted by Asia IoT Business Platform, 51% of organisations in Southeast Asia said they still need to better understand the benefits of IoT. Only 8% have implemented IoT, though the benefits are not yet clear to those organisations.
“That’s the thing about IoT – on adoption, companies begin to track and collect data which they never did before,” Irza Suprapto, director at Asia IoT Business Platform, told Computer Weekly. “It’ll take time to collect sufficient and correct data to set a benchmark to which future improvements or declines can be measured against.”
While companies are sizing up IoT and its business benefits, the tech industry needs to address two main bugbears that are hindering adoption – the high cost of upgrading legacy systems to support IoT implementations and the security of IoT devices.
Irza said the cost issue is being resolved by partnerships and business models between connectivity providers, technology suppliers and enterprises.
For example, a technology supplier in the region had offered to install digital advertising signage at a mall for free in exchange for a share of ad revenues. This reduced the capital expenditure for the mall owner, while providing the supplier an income stream beyond its traditional business.
Read more about IoT and AI in APAC
- A Boston-based supplier of artificial intelligence tools is investing around $11m in a regional office and research outfit in Singapore.
- Over a quarter of Aussies are now ready to use an internet-connected device, such as a virtual assistant or connected fridge, to make payments on their behalf.
- Malaysia is exploring the use of IoT for agriculture in ASEAN, driven by collaboration between government and the private sector.
- Nearly half of Asia-Pacific manufacturers will have fully connected factories by 2022.
From an IoT connectivity perspective, network operators such as UnaBiz, which runs a Sigfox-based low-power wide area network (LPWAN), has set its network subscription fees in Singapore at just S$1 (US$0.71) per device per month, which comes with a data plan for up to 140 messages per day.
In Taiwan, where Unabiz operates a similar network, the monthly subscription fee starts from NT$2 (US$0.07) per device for 60 messages per day.
The low cost of LPWAN connectivity offered by the likes of UnaBiz is just one part of the IoT equation. For one, the IoT market is still fragmented with many suppliers sending out different messages to the market around a handful of common use cases such as predictive maintenance and asset tracking.
“There are still limited use cases – we need tens of thousands of use cases, not one or two,” industry analyst Charles Reed Anderson said at IoT Asia in March 2017.
As for IoT security, businesses can adopt measures to prevent their IoT devices from becoming the weakest link in their IT infrastructure while IoT security standards are being worked out.
Speaking at an industry event in February 2017 this year, Hugh Ujhazy, IDC Asia-Pacific’s associate vice-president of IoT and telecoms, said enterprises can use accelerometers to track movement of IoT devices, encrypt data at rest and in-motion, and constantly profile the connection between an IoT device and its gateway to pick up anomalies.
A longer-term solution to the IoT security conundrum might be to build IoT devices that are capable of being defended. Thomas Dullien, a reverse engineer and vulnerability researcher at Google, painted a future where all system code and firmware hashes of IoT devices are part of a public ledger.
“Device and software makers make their software publicly visible to avow that they have written and stand by their software. The hashes should not be updatable, so there’s no way of messing around with integrity checks,” he said at Black Hat Asia 2017 in Singapore.
The rise of AI
Closely linked to IoT adoption is the use of AI and machine learning to make sense of streams of data being generated by IoT devices and consumers to make decisions, sometimes on behalf of organisations.
This could be pre-approving loans using chatbots in the case of Malaysia’s RHB Bank to improve customer service, or automatically shutting down a consistently faulty network device to alleviate the need to perform manual tasks in IT workrooms.
Across the region, countries that are facing rising labour costs and the need to raise dwindling productivity have been the strongest proponents of AI and machine learning.
Singapore, for example, has earmarked AI as one of the key areas in its industry transformation blueprint that includes building a library of audio and text files that enables applications such as chatbots to understand local accents.
The country’s Infocomm and Media Development Authority (IMDA) is also starting an AI apprenticeship programme to train 200 professionals over the next three years, with trainees able to hone their skills through attachments at partner organisations and AI courses conducted by educational institutions.
While governments across the region have been driving AI developments, the private sector is not standing still.
In May 2017, Huawei and Keppel said they were testing the use of AI to improve datacentre operations and energy efficiency at a reference site in Singapore.
Australia’s NAB is also testing a digital virtual banker service for business customers who can expect instant responses to more than 200 common queries its human employees have traditionally had to address.
Much of the developments in AI today, however, are considered narrow AI, which is only apt at performing specific tasks, such as recognising images. Given that the technology is still far from reaching a stage where it can take on general intellectual tasks like a human being can, fears about people losing their jobs remain unfounded, for now.
A Capgemini survey of almost 1,000 organisations that are implementing AI revealed that AI had generated new roles for 83% of respondents. Among those that had deployed AI at scale, 63% said no job had been axed.
That does not mean that workers should stand still and not strive to improve their skills, however, because doing so would be detrimental to their careers and their country’s economy.
