Towards a Smart Nation – Working with bots and AI
Singapore, February 27, 2018
Be mindful of unintended consequences of well-intentioned innovations. This was one of the main messages that came across in a keynote presentation by Dr Vivienne Ming, co-founder and managing partner of Socos, during a workshop organised by NCS at the recent Emtech Asia 2018.
Dr Ming predicted that within the next 20-30 years, many of the technologies that are now purely in the theoretical and academic realm are going to fundamentally change the definition of what it means to be human. And this is something that is worth thinking about “right now” because well-intentioned changes can sometimes make matters worse, and many successful innovations can go on to have the exact opposite impact of what people expect.
An example is the creation of ephemeral social networks like Snapchat which have no mappable history. This may be one way to protect reputations by erasing traces of what has been said online. However, studies have shown it to have the exact opposite effect: Because users believe that the messages sent out will somehow have lesser repercussions, they tend to over share, and this is causing them even greater reputational damage.
Looking to the future, Dr Ming said she is very sure that in the next 20-30 years, there will be a substantial increase in the early onset of dementia worldwide. The reason for this is the over-reliance on automated navigation system, which can lead to cognitive decline.
The acid test for a technology is, “when we turn it off, we should be better than when we turned it on”. “Are we making things better for people or are we slowly draining the creative potential of the people that we have?”
It is important, when applying technologies like AI, to solve a problem at the human level first, and then bring in the technology, said Dr Ming.
This point was reiterated by Dr Tan Kah Han, Head, Product Research and Development, NCS, in his presentation on “AI and the Final Frontiers of Digital Transformation”.
“You really have to understand the problem you are trying to solve,” he said, illustrating this with examples of how NCS is applying AI across various industries.
In healthcare, for example, AI is helping to streamline the patient journey in a smart hospital, allowing them to be discharged more quickly. In the area of public safety, sensemaking across multiple events with surveillance and video analytics helps enhance situational awareness, enabling Home Team officers to react faster to incidents.
Exploring the potential of Artificial Intelligence and Machine Learning in government digital services
The role of AI and ML in the public sector was discussed in greater depth during the roundtable segment of the workshop, where participants exchanged views on whether these capabilities could be harnessed to enhance governments’ digital services.
One of the discussion groups pointed out that the government is sitting on very rich sources of data. But the challenge is – how to bring the silos of data together for a more integrated offering, how to share the data and make better sense of it.
Take URA for example. The agency digitalises and shares a lot of its data with other agencies, and also creates joint labs with industry and partner agencies to work on projects.
In regulating accommodation for foreign workers, it looks at datasets such as electricity and water consumption to predict the location of illegal housing.
It also uses transportation data to do planning. For example, with 6 million data points from the use of cash cards daily, it is able to analyse where people work and how many trains they need to take to get to their workplace. This helps the agency to plan towns such that it brings jobs to the people and not people to the jobs.
However, there are still some issues that have to be addressed with the use of data. These include the need to look at the data privacy issue, to determine how accurate the data is, and to put in place systems to update the data and prevent the misuse of data.
Another discussion group at the workshop highlighted the importance of educating all stakeholders on the benefits of AI/ML adoption.
It noted that in this respect, the government has become more responsive in recent years. When rolling out digital services, there is a greater effort to engage with citizens, form focus groups and discuss different scenarios.
However, there are still a few issues to address in order to improve the use of AI/ML in digital services. One of these is the need to adopt common rules for structured data, and to be cognizant of cultural considerations such as the sentiments and views of citizens.
Exploring possibilities with Robotic Process Automation and chatbots
A third group at the workshop discussed robotic process automation (RPA) and chatbots, and the “limitless possibilities” that these present for enhancing the customer experience, and the challenges/limitations that the industry is facing in implementing these technologies.
One of the challenges with regards to chatbot implementation is that the technology is just off the first half of the hype cycle, and the industry is just coming to grips with its limitations. For example, there could be a lack of cultural and language sensitivity in chatbots, which may be evident especially in a multilingual country like Singapore.
Another obstacle in chatbot implementation could also be due to the fact that targets are being set too high. For example, Ask Jamie, the public sector chatbot, is more an FAQ engine than an actual chatbot.
The group concluded that if users are expecting “artificial humans”, they will be disappointed because the technology is not quite there yet - chatbots still do not have conversational intelligence.
The NCS workshop was moderated by P. Ramakrishna, Deputy Chief Executive Officer of CIO Academy Asia. Facilitators included Dr Ming, Dr Virginia Cha, Adjunct Professor of Technology Entrepreneurship, INSEAD and Dr Desai Arcot Narasimhalu, Founder and Managing Innovator, SImha Innovations.