Singapore’s AI construction boom faces data and governance challenges

Singapore is fast positioning itself as a leader in innovation and digital transformation in Asia and has made huge strides in recent years to bring AI front and centre of cutting-edge industry developments.

In the construction industry in particular, AI has already demonstrated considerable potential to streamline workflows, improve safety and productivity, and ultimately drive the industry’s sustainability. However, AI’s transformative potential is limited in Singapore by challenges around data access and governance, and skills gaps.

These were the main takeaways from a recent roundtable held in Pinsent Masons’ Singapore office, where a group of industry stakeholders gathered to share their views on how well Singapore’s construction industry is already leveraging AI, as well as what challenges persist for a country which is home to one of the most advanced AI legal and policy frameworks in the world.

Regulation and policy drive

Singapore has taken a largely ‘soft law’ approach to regulation in the AI space to date, publishing frameworks and guidance for voluntary compliance.

The country first introduced a model AI governance framework in 2019 to address ethical and governance issues related to AI deployment, which was updated in 2020. In May 2024, the government released a model AI governance framework for generative AI.

Most recently, in January 2026, in a world first Singapore unveiled a model AI governance framework for agentic AI that is capable of autonomously planning, reasoning, and acting to achieve objectives.

The government has also established a National AI Council responsible for centralising and accelerating its research and development, testing, deploying and scaling of AI solutions.

Several recent changes to Singapore’s labour and foreign worker policies are also noteworthy. The government recently introduced a new, longer work permit for AI and technology professionals aimed at attracting top foreign talent to Singapore’s digital industries. The new track, which will take effect from January 2027, will have a five-year duration rather than the two-year span of the old tech pass, bringing it into alignment with broader work permits for other high salary workers in the country.

Data challenges

While these developments have helped nurture substantial appetite for AI adoption in Singapore’s construction industry, the roundtable revealed there is still considerable misunderstanding surrounding what specific advantages AI technologies can bring to the industry.

For instance, like many other countries, certain BIM and advanced project management tools have been used frequently in recent years to streamline construction activities and prevent design clashes across the country’s construction industry. These predate and should not be confused with more recently developed generative AI-powered tools that have capabilities to learn from data and improve processes and outcomes in real time.

AI’s effectiveness in Singapore’s construction industry also depends on access to large, reliable, and diverse datasets. Despite the ubiquity of major tech and IT providers in the country, there is still a distinct lack of construction-specific data to offer genuinely transformative AI tools.

Engineering and construction firms often hold valuable data but are reluctant to share it due to confidentiality concerns, which limits the prospects for meaningful collaboration across the industry. Many firms also favour internal tools or are developing their own proprietary AI models, but even larger firms don’t hold enough data globally to undertake meaningful testing and training.

Governance and risk

Artificial intelligence is no longer a future concept in Singapore’s built environment. It is already reshaping how we design, build, operate and manage risk across projects in the country, but there are still critical challenges ahead.

From a governance perspective, there is still a need for clearer processes to control and oversee AI use across projects and supply chains. Construction is a safety-critical industry and, as in other jurisdictions, a lack of digital literacy makes these changes as challenging in Singapore as anywhere else.

Questions around data quality, data-sharing, accountability, contractual allocation of risk and intellectual property cannot be treated as afterthoughts. In an industry built on responsibility and long-term value, having clear governance in place, including project controls for AI-related design and construction risk, will be essential.

From an insurance perspective it was interesting to hear how risk is handled in relation to AI and that errors exist regardless of whether a design is produced by humans or by AI.

Some stakeholders suggested that AI may even improve risk assessment and pricing since insurers already possess large datasets and see considerable potential to deploy AI internally to make their assessments more accurate. In this sense, rather than increasing liability, AI-enabled tools may reduce risk by identifying issues early on.

There is clearly still more work to be done to understand better how AI affects responsibility and liability, but, for now at least, insurers appear neutral in viewing AI as increasing risk in Singapore’s construction industry.

Arguably, one of the most important takeaways from the roundtable is that AI adoption in the built environment must remain human-centred. Engineers, architects, contractors, owners and advisors will continue to exercise judgment, experience and professional responsibility. AI should be seen as a tool that supports better outcomes – safer projects, more sustainable assets and fewer disputes – not as an end in itself, but as an enabler and catalyst for future and sustainable improvements.