Buy a new TV, plug it in and turn it on. Immediately you can watch your favourite programmes.
Artificial intelligence (AI) is anything but this easy, said Chua Chwee Koh, chief of group technology and operations at security company Certis.
Using autonomous vehicles (AV) as an analogy, he said AI comprises a large part of the AVs but it is difficult to put them on the road partly because the public expects autonomous cars to be safer than real cars.
So while AVs have been developed for a long time, mass implementation will take a while, he added. “That’s why the world’s largest car makers like Ford and General Motor have delayed the launch of AV cars.”
A successful AI service or solution, he pointed out, must be operationalised to fit company’s challenges. It requires a team of non-tech and tech specialists to create such solutions, he added.
The first step is to understand the business challenges, not try to fit AI into a solution. Design thinking maybe introduced to understand users, challenge assumptions, redefine problems and create innovative solutions to prototype and test.
This is an iterative process but it will help break down the silos in organisations, bring together different departments to improve synergy and provide better experiences.
This approach is called ops-tech, said Chua. Certis has introduced it into its Security+ service which combines security, facilities management and customer service into a holistic security service. This has been rolled out in major shopping malls and infrastructure facilities.
Said Chua: “The difficulty for using AI in a business solution is to get all these components together into a service. Easy to say, tough to execute,” he said at a presentation during Nvidia’s AI Innovation Day held at the Kent Ridge Guild House in Singapore on December 4.
Certis is also applying its AI-based services in areas to improve workplace safety, identify different species of mosquitoes and predict energy consumption.
Speakers at the one-day Nvidia AI Innovation Day discussed the importance of AI platforms to make it easy for developers to develop AI solutions.
Dr Ettikan Karuppiah, Nvidia’s director for developers ecosystem for Southeast Asia, Australia and New Zealand, said that the company has made available pre-trained models for developers to create their own deep learning projects.
Pre-trained models come with data collected elsewhere. However, developers can replace the data with their own information to re-train the models, making it easier and faster to deploy their own deep learning systems.
Before starting any development, Dr Karuppiah highlighted that organisations should among other things identify the key metrics whether it be saving money, increasing efficiency or ensuring people safety.
It is also necessary to run a proof of value concept to ensure that the key metrics are on track, he added. Once this is proven, then deployments should start small with the view to scaling upwards, he added.
Wherefore Singapore’s AI vision? It is to ensure that by 2030, Singapore will be a key AI hub, said Dr Chng Zhenzhi, director of the National AI Office, from the Smart Nation and Digital Government Office.
She elaborated on the Republic’s AI vision which comprises a two-step strategy. First is to develop impactful solutions in five sectors including transport and logistics, healthcare and education.
Second is to grow the AI ecosystem including ensuring sufficient talent, establishing a progressive and trusted environment and advancing collaboration among companies and research institutions.
