In the AI era, here’s how tech hub Singapore partners companies for success

Despite Southeast Asia’s strong push towards adoption of Artificial Intelligence (AI), a recent IBM study revealed that only four per cent of ASEAN organisations have reached the transformative stage of AI readiness.

Conducted by digital research and advisory company, Ecosystm, for IBM, the study’s findings revealed a stark contrast between enthusiasm for AI and actual execution or adoption, underscoring the challenges businesses face in reaping the full benefits of AI.

IBM’s ASEAN General Manager and Technology Leader, Catherine Lian notes: “To power ahead in an AI era, organisations need to focus on a robust data strategy, accessibility and AI skills. They must also proritise data governance and compliance especially with regulators worldwide looking to regulate AI technology to ensure citizens, businesses, and government agencies are protected from potential risks.”

Catherine contends Singapore has made significant strides in supporting companies to adopt, deploy or drive AI innovation here. She adds that Singapore’s focus on talent and education, partnerships, data, and digital infrastructure, along with active collaborations with the international community, have created an environment where AI is trusted by its citizens. 
 

She says: “Trust and transparency are vital to fostering technology that will both augment and transform the way we live and work.” 

She shares how government support and access to partners in Singapore has enabled IBM to successfully implement its AI strategy, including integrating sustainability into its projects:
 

1. From IBM’s perspective, what makes Singapore an attractive destination for AI development and deployment?

Singapore has taken the right steps in laying the policy foundation for AI development and adoption at a national level. The market-friendly and multi-stakeholder approach undertaken by the Infocomm Media Development Authority (IMDA) is a reinforcement of the collaborative approach the government has undertaken in furtherance of AI in Singapore. This is an approach that IBM is supportive of and a part of.

For instance, we recently collaborated with AI Verify Foundation as a design partner and contributor to Project Moonshot, an easy-to-use testing toolkit platform designed to address security and safety challenges often associated with the use of large language models (LLM). The provision of this new tool is significant as it would help developers and data scientists test their LLM applications against a baseline of risks, thereby seeing the direct value of AI, rather than just the model that providers see the value.

We are also working with AI Singapore to test the first LLM with Southeast Asian context —SEA-LION LLM—using our data and AI platform, watsonx, and have also embarked on technical exchanges and knowledge sharing to enhance the model.
 

2. IBM has been in Singapore for decades. How did this factor into the company’s decision to launch its latest round of AI initiatives here?

IBM has been in Singapore for 71 years since we first established a presence here in 1953! As part of our continued commitment, we recently announced our intent to establish a new AI research and innovation centre with the National University of Singapore (NUS). Based at the NUS School of Computing, the centre will feature the first full-stack AI infrastructure system installed on a university campus in Asia-Pacific by IBM.

This partnership demonstrates the strong commitment by Singapore’s research and innovation ecosystem to collaborate with industry leaders, such as IBM, to realise Singapore’s National AI Strategy. Equipped with full AI infrastructure and supported by the strong research expertise of NUS and IBM, the centre will aim to serve as an innovation platform for Singapore agencies, academic and research institutions, as well as companies to jointly conduct cutting-edge AI research with significant translational potential.
 

3. What are some of IBM’s key AI projects, successful use cases or collaborations in Singapore and Southeast Asia?

IBM is a longstanding partner with the private and public sectors in Singapore and the Association of Southeast Asian Nations (ASEAN) to drive AI use cases and responsible AI adoption. 

For instance, in Singapore, IBM plays a pivotal role in implementing our data and AI solutions, watsonx.data and watsonx.ai solutions, to enable telecommunications company Starhub to securely centralise its network-related data. This centralised repository facilitates streamlined big data processing and analysis, laying a solid foundation for future endeavours such as data hub establishment and data monetisation for B2B and B2B2C businesses.

Last year, we also announced a collaboration with the Singapore Civil Defence Force (SCDF), supported by IMDA and Starhub. Through deploying AI and 5G-connected Augmented Reality (AR) smart glasses, SCDF automated its equipment inspection process to ensure frontliners’ operational readiness.

We also play a key role in developing frameworks and best practices for the ethical use of AI. IBM is currently a premier member who will set strategic directions and development roadmap of responsible AI for the AI Verify Foundation.  Under Singapore’s leadership, ASEAN in February 2024 also published a framework for artificial intelligence governance and ethics. IBM is one of the companies consulted for this framework.
 

4. How is IBM integrating environmental sustainability into its AI projects?

Sustainability matters—for businesses, for society, and for everyone. In Singapore, we continue to drive energy efficiency through innovation and have yielded positive outcomes.

Take for instance the AI system at the proposed AI research and innovation centre with NUS. The Artificial Intelligence Unit (AIU) accelerator is IBM’s first complete system-on-a-chip for running deep learning models faster and more efficiently than a general-purpose CPU. The specialised computer chip for AI enables over four times the power efficiency against commonly used GPUs, capable of more energy-efficient AI workloads.

There are other ways that IBM is innovating to minimise the environmental impact of AI. For instance, the amount of power required to train a single GPT-3-sized model is equivalent to the yearly electricity consumption of over 1,000 American households¹. IBM’s Granite 13B model delivers performance that is just as competitive with a significantly lower carbon footprint and is smaller in size compared to massive 100 billion or 1 trillion (or more) parameter models.

Footnotes: