Facebook and Amazon’s AI-selected ads show how businesses can tailor products to consumers in future

Many people are worried that the increased automation AI brings will make things impersonal.

However, Dr Nigel Duffy, artificial intelligence leader in global innovation at EY, one of the “Big Four” accounting firms, who is also speaking at IES, rejects this view and believes the use of the technology “will make the world more humane”.

He says: “It will give us freedoms that we cannot yet imagine. It has the potential to improve society broadly in terms of health care, housing, city planning and transport.

“Roles and tasks will be enabled by AI – meaning employees and organisations will operate with more information and intelligence to get tasks done better and faster.

“By reducing repetitive work, AI will allow individuals to spend more time and energy on building and managing relationships, give medical professionals the tools to connect with patients better, and provide enterprises with the time and resources needed to develop products unique to individual needs.”

Enhanced efficiency and lower costs

China’s second biggest e-commerce company JD.com, which serves more than 250 million active shoppers, has used AI to carve a niche for itself.

The company’s research into AI includes speech and audio, natural language and semantics, knowledge graph, computer vision and machine learning.

“The current technology is already making a significant impact in helping companies to improve efficiency, reduce costs and enhance experience,” Dr Xiaodong He, technical vice-president and deputy managing director at JD AI Research, says.

He, who leads the JD.com Deep Learning, Speech and Language Lab, will also be speaking at IES.

One of JD.com’s AI-powered technologies is a robot that combines a chatbot’s functionality with customer service from a human agent.

Using technologies such as natural language processing and machine learning, the chatbot is designed to recognise seven emotions, provide personalised feedback according to changes in a customer’s mood and seamlessly transfer the conversation to a human agent as required.

“To put it in perspective, during our Singles’ Day shopping promotion, [for the world’s biggest online shopping festival held on November 11 each year] more than 16.3 million customer service inquiries were handled by our smart customer service robot, of which 90 per cent were handled entirely by the robot,” He says.

Recognising humans and snacks

Besides functioning as a chatbot, another AI technique that shows great promise is facial recognition, which can be applied in fashion stores to provide personalised recommendations for shoppers.

“It has already been adopted in the legal sector, providing more reliable results than other traditional methods,” Chiribella says.

“The same principle sees it applied to the medical sector to speed up diagnoses and process complex sets of data from multiple patients and multiple hospitals.”

Last month, JD.com announced a collaboration with Hong Kong’s Fung Retailing, which has started a second test of an AI-powered solution at two of its Circle K convenience stores.

The technology completes the checkout process in four seconds.

There are three steps to checking out: place products on the smart counter; press the AI Recognition button on the cashier screen to have the products scanned; and pay with an Octopus stored-value card.

Making sense of ‘big data’

Away from the retail sector, many people already have a personal relationship with Alexa and Siri, the computerised smart voice assistants embedded in our home technology.

Experts believe that enhanced interaction between humans and machines is another important direction for AI.

“In the future, we will improve the interface with our devices, making them more intuitive and more like the interaction we have with humans,” Chiribella says.

“It is not entirely science fiction to imagine that there will be applications like the e-friend, capable of entertaining the user with simple, human-like conversations.”

However, perhaps the most important application of AI is in making sense of big data – very large sets of data that may be analysed computationally to reveal patterns, trends, and associations, especially relating to human behaviour and interactions.

“Businesses will have to use AI techniques to analyse the preferences of their customers, to monitor their reactions to new products, or to assess the success of an advertising campaign,” Chiribella says.

“The same applies to organisations and institutions when it comes to evaluating the success of new polices or to understand trends in public opinion.”

Firms must get ready for AI

Although AI technology is no longer new, its widespread adoption by Hong Kong businesses is far into the future.

JD.com’s He says interest in AI is high, but so far implementation is low because there is a lack of understanding about the technology and only a small pool of AI practitioners.

However, the time is right to prepare for adopting the technology, He says.

“Businesses have to plan strategically; they need to learn what AI can do in their business transformation.

“They have to clarify the demands of the business and formulate an overall plan based on goals, and make sure AI will be smoothly integrated in their organisation.”

Duffy says companies should avoid focusing too much on the cost efficiencies that AI provides.

“Businesses must ask themselves how much more value they can deliver through more intelligent use of AI,” he says.

“This means identifying the problems that need to be solved first, and then creating solutions through AI that are tailored accordingly.”

Duffy believes that the successful integration of AI will not be determined by the technology alone. “[It] will be largely determined by talent, leadership, cross-functional collaboration, communication and trust,” he says.

“Leaders of organisations should invest in hiring and developing their people to ensure that they have the right skill sets to maximise the potential of AI.”

Lack of trust in technology

For most companies, one big concern is how to know that their AI system is doing what it is supposed to do.

“To establish greater trust in AI, CEOs and other business leaders must validate the technology is behaving how it is supposed to,” Duffy says.

“Transparency, bias and diversity are all important in establishing trust in AI.”

JD.com’s He adds: “To achieve and sustain this trust, an enterprise must understand, govern, fine-tune and protect all of the components embedded within and around the AI system – AI’s performance must show precise and consistent outcomes.

“The potential is huge and includes analysing big data, revealing hidden patterns, engineering new materials and designing networks of sensors with enhanced capabilities.

“But, the applications to AI are still in their early days and it is premature to assess their impact.”

Editor’s note: The Internet Economy Summit, which was being held on 15 and 16 April at Hong Kong Convention and Exhibition Centre, in Wan Chai, had Dr Nigel Duffy to speak at the Visionary Forum on the first day while Dr Xiaodong He and Giulio Chiribella spoke at the AI Forum on the second day. “Digital Economy – Redefines Our Future” was the theme of the technology flagship event – organised jointly by the Hong Kong government and Cyberport, the city’s business park formed by about 1,000 digital technology companies.