AI sustainability top of mind for Singapore public sector – New report

By Si Ying Thian

More than two-thirds of Singapore public sector are concerned about the increased energy consumption with using AI, but their understanding of newer energy-efficient tech is relatively low, according to a new Pure Storage survey.

77% of the Singapore public sector organisations surveyed highlighted increased energy consumption from using AI as a top concern, according to Pure Storage's new report. Image: Canva.

Between the national push for artificial intelligence (AI) and the energy-intensive nature of AI, the Singapore public sector is facing a delicate balancing act of deploying AI while meeting its sustainability targets.


The country’s public sector has set an ambitious target of achieving net-zero emissions around 2045 - five years ahead of the broader national target of net-zero by 2050. At the same time, Singapore’s public sector is also the most AI-friendly according to a report by Oxford Insights.


This tension was evident in findings from Pure Storage’s new report Drivers of Change: Meeting the Energy and Data Demands of AI Adoption in Singapore’s Public Sector, which highlighted the key challenges faced by Singapore’s public agencies in pursuing AI.

AI vs sustainability: Is there a trade-off?


Of the 40 government leaders surveyed, close to all (97%) are reportedly already deploying or trialing AI in their operations.


However, the increased energy consumption from using AI is a top concern for 77% of them, as it impacts their ability to meet ESG/sustainability targets.


In the report, a respondent coming from an essential services provider even emphasised the “super critical” need for them to meet the sustainability metrics every year.


The public sector currently adopts multiple approaches to mitigate the additional energy consumption of AI. The top three strategies shared were to invest in more energy-efficient hardware, engage in energy-saving practices, and reduce energy usage in other areas.


The devil is in the details. While respondents pointed to the need to invest in energy-efficient hardware, they have a rather limited understanding of newer energy-efficient tech when asked about those.


Recently at the Asia Tech x Summit 2024, Singapore’s Infocomm Media Development Authority (IMDA) announced the Green Data Centre Roadmap, which laid out its plans to partner with the industry tap on the latest energy-efficient tech.

AI deployment has not been all smooth sailing


A significant majority of respondents (83%) reported a need for a complete or substantial overhaul of IT infrastructure to deploy AI in its operations, which suggests a lack of AI readiness.


Within IT infrastructure, organisations underestimated the data storage and management, computing power, and energy consumption requirements needed to adopt AI at scale.


87% of those surveyed pointed to flash storage technology as one of the possible solutions to optimise their IT infrastructure for energy efficiency.


In a previous interview with GovInsider, Pure Storage’s Hock Leng Chua shared that all-flash storage has a higher storage density than off-the-shelf solid-state disks, which can help organisations reduce their total power usage by between five to 10 times.


A higher storage density means that it can store a larger quantity of data with less space.


Other barriers that reportedly stop organisations from deploying AI include the lack of in-house skill sets, especially so in a small market like Singapore, and the high investments needed.


Conducted from February to April this year, the report surveyed leaders from Singapore’s public sector, including government agencies, essential services, health and education. It was supplemented with additional qualitative insights from semi-structured interviews.


The leaders were either IT decision makers or have the authority to influence IT purchasing decisions.


To find out more about the state of AI adoption in Singapore's public sector, challenges and strategies for mitigation, you can download Pure Storage's whitepaper by clicking the button below:

This article was produced in partnership with Pure Storage.