Ever since ChatGPT vowed the masses in December 2022 and triggered the minds of a million entrepreneurs, Artificial Intelligence (AI) has quickly moved from being a term in academia and science fiction to being the central focus of many public conversations. Every government, enterprise, and individual has started to think about the accessibility and impact of AI with optimism around productivity and hitherto unavailable capabilities, and pessimism about disruptions causing job losses and an inability to upskill to benefit from AI. Practitioners in every sector have wondered about the impact of AI and how it could be leveraged towards greater economic and social value. The sectors of intelligent transportation and smart mobility have always been eager adopters of technology and AI is no exception. But is AI ready to be applied to these sectors, and are these sectors ready for AI? This year’s edition of the annual ITS Singapore Summit, organised by the Intelligent Transportation Society of Singapore (ITSS) in July 2024, delved into these ideas, specifically as they apply to Singapore – a nation that has been at the forefront of intelligent transportation and smart mobility. In this piece, we consider how AI might apply to the intelligent transportation and mobility sectors and why Singapore might be a great place for it to happen.
The recent explosion of AI in public discourse has largely been driven by Generative AI that has garnered a lot of attention due to capabilities such as creating, interpreting and summarising text, images and video. However, the journey of AI started in the 1950s and has passed through technologies such as machine learning, data analytics, deep learning, and now, generative AI. Over the years, Intelligent Transport Systems (ITS) have adopted these technologies to improve performance and productivity with many of these systems being deployed in Singapore.
ITS deployments regularly include sensors and systems that embed intelligence derived through artificial intelligence, machine learning and data analytics. These include the following:
Without a doubt, these systems have delivered localised benefits towards a higher quality of service and embedded intelligence.
Looking ahead, Generative AI opens new avenues to target both mainstream and long-tail use cases. Some of these will extend existing systems and some will be completely new applications that are enabled as the technology matures.
The obvious uses for Generative AI are in automating tedious work such as preparing incident reports for cross-agency sharing and collaboration. The more nuanced use cases would extend to personalised journey planning and aiding inclusivity for differently abled citizens and tourists who speak a different language.
AI will impact all aspects of the ITS operations value chain: information gathering, information analysis and sense-making, and decision support and service delivery. Current systems already use smarter sensors that process data on the edge and more becomes possible as hardware becomes more capable. AI can also greatly improve information analysis and sense-making in systems, for example, by analysing patterns and anomalies to do automatic incident detection. These can be further enhanced by natural language processing on large datasets that will even parse through social media feeds to find incidents and frustrations reported by citizens. AI also has applications in quickly generating, simulating and evaluating response plan scenarios to provide decision support to service controllers when an incident happens.
Beyond operations, large-scale processing of real-time data feeds will enable deeper insights about commuter and driver behaviour, allowing planners and authorities to address the challenges of urbanisation.
Over the past few years, an increasing amount of data has been collected and generated within intelligent transport and smart mobility applications. These might include private data of citizens or information that citizens may want to keep private. Data management, including data privacy, data safety and limiting access to authorised individuals remains an important requirement. In Singapore, the Personal Data Protection Act (PDPA) provides guidance and regulations that enterprises must follow where private data is concerned.
In addition, there is a need to ensure that the capabilities enabled by AI are used responsibly. This will necessitate oversight and governance for both data and AI. Countries across the world have tried to approach this issue from different angles. The AI Act in the EU has come under criticism for demanding much more than the average startup might be able to deliver. Governments will need to ensure that the frameworks strive to find the balance between allowing innovation and exploration, and still protecting the data and rights of the citizens. Moreover, the use of AI in safety-critical applications, e.g., control of vehicles or traffic signals will need strong frameworks to ensure that the deployments are safe or provided with additional safeguards to protect the citizens – both inside and outside vehicles. For this reason, a cybersecurity assessment framework for AI should consider security from the start, and across planning, data collection, training, evaluation and operation.
It is clear that the next phase of AI adoption has the potential for greater impact but needs everyone to be more involved. Practitioners and thought leaders regularly stress the need to include everyone in an “AI Transformation” rather than just making some systems smarter. This means providing training for AI users as well as developers of applications and AI systems. It also means adopting a “no one is left behind” mindset to ensure that everyone benefits.
In this area, we are very pleased to see the investment, initiatives and support that Singapore is providing enterprises. This includes support for training (including free courses for everyone to understand how AI would impact them) and talent development for deep-tech professionals. In addition to such resources, AI Singapore offers full-time apprenticeships for professionals wanting to adopt AI in their careers and support for proofs of concept and full systems. Industries can work with AI Singapore or with Institutes of Higher Learning to build AI systems that can either become a part of their products or help them solve internal problems.
The work done by Govtech to define the parameters of Government on Commercial Cloud (GCC) has also meant that for a vast array of applications, suppliers to government agencies are able to use cloud providers such as Amazon Web Service, Microsoft Azure and Google Cloud Platform to deliver AI solutions. Cloud providers have large infrastructures within data centres in Singapore (thereby meeting data residency requirements) and can flexibly support the computing and storage demands of these cutting-edge applications, along with developer toolkits (APIs, etc.) that simplify the use and integration of AI within systems. Singapore has also made advances in green energy to tackle the anticipated demand for energy that such systems might need.
Most uses of AI in transportation might not suffer from concerns like deepfakes and scams but trust in AI will come from knowing that the systems are transparent, fair and explainable. Where relevant, a certification or review of the systems might encourage greater trust. Towards this, Singapore has moved swiftly, yet gingerly, to outline a national strategy for AI (all the way since 2017) and has developed safeguards to help enterprise and public sector users.
Finally, specifically towards transportation, the Singapore Land Transport Authority (LTA) has also established an AI Office to oversee the use of AI and to address fairness and privacy concerns. In addition, the LTA has announced that they are establishing a Road Innovation Lab that will allow collaboration between the public and private sectors to jointly exploit and benefit from the development and deployment of AI in transportation.
In this year, there has been a lot of talk about AI and deployments of intelligent transport systems have continued in full earnest. When looking back and looking ahead, we feel:
As the Intelligent Transport Society of Singapore, we are optimistic about the use of AI in our sector and look forward to our members working closely with each other to deliver this new future in Singapore and in other markets.
The Intelligent Transportation Society of Singapore (ITSS) was founded in 2001 with the aim of bringing together the professional interests of those in public and private organizations, practitioners, academics, and researchers, related to ITS, and to create opportunities for networking and interaction. Today, the vision of ITSS is to be the champion for Singapore's ITS and mobility community, towards transportation that enhances quality of life.
Our mission:
