Professor Anthony Tung, AI Lead for Urban Sustainability at the NUS AI Institute
As companies around the world continue to explore artificial intelligence (AI) solutions, organisations often face several hurdles in successfully adopting AI technologies.
Professor Anthony Tung, AI Lead for Urban Sustainability at the NUS AI Institute in Singapore, and a faculty member in the Department of Computer Science at NUS School of Computing in an exclusive interview with iTNews Asia said these challenges stem from factors such as a lack of AI understanding, limited talent, data constraints, and insufficient computational resources.
“Many companies don’t fully understand what AI can or cannot achieve, leading to a lack of motivation to explore its potential,” said Professor Tung. “They also struggle to find skilled talent, as AI expertise is in high demand across industries.”
In addition to limited manpower, many companies face a “small data” problem. Instead of having clean, labelled datasets for training AI models, they often collect data only for operational needs, which isn’t ideal for AI and analytics.
“Proper data management is a prerequisite for AI adoption. Without it, even the most advanced AI models will struggle to perform effectively,” Professor Tung explained.
He also highlighted the financial burden that comes with acquiring the computational resources necessary to train and maintain AI models.
Ongoing AI projects in NUS
Professor Tung said Singapore’s National University of Singapore (NUS) has made significant strides in AI research and application.
Apache SINGA, an open-source distributed deep learning system is one of the scalable AI systems for large scale data analytics.
The other initiative, RETINA, from collaboration between NUS and the National Healthcare Group Polyclinics improves the screening of retinal images for medical conditions. The project has already deployed a prototype in multiple healthcare centres, he added.
NUS also supports organisations to evaluate their AI readiness with the AI Maturity Model, which tracks progress from initial awareness to fully integrated AI systems.
When it comes to adopting AI, Professor Tung stressed the importance of having a comprehensive AI roadmap that incorporates business objectives, data quality, infrastructure readiness, and workforce development. “AI projects should not be seen as standalone initiatives but should be integrated into the company’s broader strategy,” he said.
For companies looking to deploy AI solutions, he recommended preparing the IT infrastructure. Organisations must assess their current hardware and software capabilities, investing in scalable cloud solutions and high-performance computing resources. This will ensure that AI workloads can be handled efficiently, he added.
Professor Tung said the NUS team is currently working on creating ‘White-box AI’, which aims to make AI systems more transparent, controllable, auditable, resilient, and antifragile.
Looking ahead, he believes that the next wave of AI innovation will focus on trustworthiness. “As AI systems become more integrated into our daily lives, transparency, controllability, and resilience will be critical in ensuring their safe and effective use”.
With continued investment in AI research, partnerships with industry leaders, and a strong ethical framework, he hopes AI can be trusted to deliver transformative, positive impact.