Artificial intelligence takes shape

AI and ML revolutionize modern businesses (image NicoElNino / iStockPhoto)

“The development of AI in Hong Kong is balanced across different vertical industries from government, banking, financial, healthcare, manufacturing to hospitality industries,” says Tracy Tsai, research vice president at Gartner.

AI is one of the major innovation and technology (I&T) development areas under the support of the HKSAR Government. In the 2018 Policy Address, the Chief Executive stated that the funding of HK$10 billion has been approved for establishing two I&T clusters. One is AI and robotics technologies. In addition, AI and chatbot functions will be introduced to the GovHK portal in 2019 to facilitate searching and access of e-Government services by the public.

More companies embrace AI to enhance customer experience, optimize business processes, or increase operational efficiencies. Joint AI research centers or labs were set up by different educational institutions and technology providers in 2018.

Areas of development

According to Tsai, AI development is mostly focused on three areas. They are computer vision applications, natural language processing (NLP)-enabled virtual customer assistants, and robot broker or fraud and risk analysis.

Gartner's Tracy Tsai

In the computer vision area, homegrown AI unicorn startup SenseTime sees a growing demand for biometric authentication in the financial sector.

“Biometric detection technology relies on sophisticated computer vision and machine learning (ML) algorithms,” says Hailong Shang, Hong Kong managing director at SenseTime. He noted many financial institutions have implemented or run pilot tests in access control or Know-Your-Customer (KYC) scenarios for identity authentication.

Virtual assistants or chatbots are also rapidly gaining ground. IDC and Hong Kong Computer Society (HKCS) noted a rise in the number of multi-lingual chatbots for enhanced customer support.

IDC's Chris Marshall

“An increasing number of companies have started to offer AI and ML service to teach chatbots in learning how to respond to different scenarios and across multiple languages for value-added services,” says Chris Marshall, associate vice president for big data and AI at IDC Asia/Pacific.

“There’s a strong interest in multilingual, NLP-enabled chatbots, as more companies look towards AI to automate customer support,” says Andy Chun, convener of artificial intelligence specialist group at the HKCS.

In the fraud or risk analysis area, SAS pointed out that AI and ML techniques are used in real time fraud detection. “They can reduce false positives and predict fraudulent transactions to safeguard banks’ reputation and customers’ asset,” says Wilson Ho, general manager of SAS Hong Kong and Financial Services at China.

Cloud-based AI development platforms and tools are also on the rise to help companies integrate AI capabilities into their businesses.

“I believe many companies prefer to develop AI by leveraging proven API services for faster idea to market cycles,” says Chun from the HKCS.

IBM and Microsoft share a similar view. According to Samson Tai, distinguished engineer and CTO at IBM Hong Kong, AI or ML model development platform on cloud is in high demand.

Microsoft sees a growth for cloud solutions and AI productivity tools driven by AI and ML initiatives. “Companies can build the next generation of smart applications where their data lives, in the intelligent cloud, on-premises, and on the intelligent edge,” says Bess Chung, cloud & enterprise product marketing manager at Microsoft Hong Kong.

Opportunities & challenges

AI will continue to transform every business in 2019. It will continue to play a role in companies’ digital transformation initiatives. AI-as-a-Service will be widely available in the market.

IDC predicts that by 2019, 40% of digital transformation initiatives will be supported by AI or cognitive capable systems, providing timely, critical insights for new operating and monetization models in Asia Pacific. The research firm also anticipates that AI-as-a-Service from major vendors will be “ubiquitous as a way for companies to experiment with AI.”

“Major vendors are making AI technologies more accessible to developers and simplifying complex ideas for people to understand and able to utilize these technologies or tools,” says IDC’s Marshall. He added that vendors’ auto ML tools lessen the complexity of developing and turning complex ML models.

On a similar note, IBM’s Tai says more robust AI services are exposed as micro services on the cloud. “They allow enterprises to integrate easily with their existing system of engagement applications.”

Tsai from Gartner has a different view. She says customization is needed to test, train, and optimize AI models based on specific customer requirements and different use cases.

Talent shortage, data quality, and the trust in AI are some of the barriers to AI development in 2019. Tapping talent from universities or overseas countries is a possible solution to tackle the AI talent problem.

Microsoft’s Chung and SenseTime’s Shang suggest companies partner with educational institutions and associations to nurture AI talent, as well as to retrain employees internally.

On a similar note, Chun from the HKCS believes many companies will tap into nearby cities for talents or outsource to companies in China or India.

Andy Chun of HKCS

AI is data-dependent. The quality, security, and transparency of data become vital for building customer trust and broader implementations on AI.

“AI needs access to lots of data and some of the most useful data is held by the government,” says IDC’s Marshall. He urged governments to be proactive in this regard.

“The danger is in ‘garbage in garbage out’– AI is only as good as the data it uses for machine learning or data analytics,” says Chun.

Similarly, Ho from SAS says “right” data is critical for a company to start using AI and ML. “It is necessary to understand how to collect and process the right data in a sizeable volume,” says Ho.

Microsoft’s Chung noted winning customer trust is the biggest market challenge as people’s daily lives are infused with AI and big data. Data breaches will shake customer trust in new technologies such as AI.

Tai from IBM says AI fairness toolkits give customers more transparency into AI. They provide customers with tools and knowledge to “integrate bias detection as they build and deploy ML models”.

Revenue generation

AI will become a major source of new revenue for companies and Hong Kong’s economy in the next few years.

“New revenue generation will take place after customer experience is optimized, which is expected to be 2020,” says Gartner’s Tsai. “I’d expect enterprises to start to use AI for new revenue stream when they acquire more data,” says IBM’s Tai.

“AI-derived business value will gradually become the dominant source over the next few years,” says Microsoft’s Chung. She cited a study conducted by IDC and Microsoft in April, which stated that by 2021, 60% of Hong Kong’s GDP will be derived from digital products such as AI, cloud, IoT, and mobility.