Singapore’s adoption of artificial intelligence (AI) could raise its potential economic output by 1.06 percent in the long run, offering a partial offset to the drag from an ageing population, although the transition could put pressure on some workers, particularly fresh graduates, the ASEAN+3 Macroeconomic Research Office (AMRO) said on Wednesday.

AMRO said in its annual consultation report on Singapore that the city state had taken an early, economy-wide approach to AI adoption, creating opportunities to boost productivity as the country faces slower labor-force growth.

However, the benefits will depend on whether businesses move beyond experimentation and integrate AI into core operations.

Under AMRO’s baseline scenario, only about half of the eventual potential-output gain is expected to be realized by 2035, rising to 0.65 percent by 2040 and 0.78 percent by 2050.

The estimates are calibrated counterfactuals rather than forecasts and focus on domestic supply-side effects.

They exclude potential gains from external demand for Singapore’s AI-related trade and investment, as well as potentially significant effects from innovation and the creation of new tasks.

Singapore’s ageing population is expected to increasingly weigh on potential growth from 2030, making productivity gains from AI more important.

AMRO said wider AI adoption could raise the productivity of a shrinking workforce and partially offset demographic constraints.

Adoption accelerating but remains uneven

AI adoption among Singapore businesses has accelerated, but the pace varies significantly by company size, said AMRO.

According to the Infocomm Media Development Authority (IMDA), AI adoption among small and medium-sized enterprises (SMEs) more than tripled to 14.5 percent in 2024 from 4.2 percent in 2023. Adoption among non-SMEs rose to 62.5 percent from 44 percent over the same period.

The Ministry of Manpower’s 2026 establishment survey showed that AI adoption was also strongly correlated with company size, with adoption at 23.9 percent among firms employing fewer than 25 people, compared with 76.4% among firms with more than 500 employees.

AMRO said the gap suggests larger companies are better positioned to invest in AI capabilities, while smaller firms continue to face barriers including implementation costs, a lack of in-house expertise and data-governance concerns.

More importantly, adoption has yet to translate into widespread integration of AI into core business processes.

Only 3.8 percent of firms surveyed by MOM had integrated AI into their core operations, while most remained at the planning or pilot stage.

Among firms already using AI, off-the-shelf generative AI tools accounted for 84 percent of adoption, while fewer than half had implemented customized or proprietary AI solutions.

AMRO said the findings suggest that simply increasing the number of firms using AI may not be enough to generate economy-wide productivity gains.

Businesses will need to redesign workflows and integrate AI into day-to-day operations to capture deeper productivity benefits, it said.

AI exposure concentrated in professional jobs

The shift could also reshape Singapore’s labor market, where the share of jobs exposed to AI is relatively high compared with other high-income economies.

AMRO estimated that occupations with a high degree of automation exposure, where AI could substitute for a substantial share of existing tasks and potentially reduce labor demand, account for 22.3 percent of employment.

The exposure is concentrated in professional occupations involving financial analysis, business administration, sales, marketing and public relations, where AI is increasingly capable of performing routine analytical and administrative tasks.

However, AMRO said exposure to AI does not necessarily mean widespread job replacement.

Many occupations consist of multiple tasks, allowing AI to automate routine activities while complementing workers in functions requiring judgement, interpersonal interaction and decision-making.

Singapore also has a relatively large share of jobs that are likely to be augmented rather than replaced by AI, particularly among women.

Business administration professionals, for example, could use AI for data processing and preliminary analysis while retaining responsibility for interpretation, decision-making, regulatory compliance and client engagement.

Firm-level evidence also points to job restructuring rather than broad-based employment losses. Among AI-adopting firms, 18.9 percent reported job redesign and 13.9% reported creating AI-related jobs, compared with 6.2% that reported reducing headcount.

Fresh graduates face adjustment pressure

Despite the limited evidence of broad-based displacement, AMRO said younger workers entering the labor market could face greater adjustment pressures.

Entry-level jobs often involve routine tasks that provide graduates with their first pathway into professional occupations.

As AI takes over some of these tasks, opportunities for young workers to gain initial experience could become more limited.

AI could also reduce the importance of geographical proximity for some service-sector jobs by lowering language barriers and allowing more work to be delivered remotely, potentially increasing competition for entry-level positions.

AMRO noted that the proportion of tertiary graduates in the labor force who were employed declined in 2025.

At the same time, Singapore’s youth unemployment rate remained at 6.6 percent in both 2024 and 2025, while the share of young people not in education, employment or training improved to 3.8 percent in 2025 from 4.1 percent in 2024.

The report stressed that these developments do not establish an AI-driven employment effect. Graduate hiring also slowed across advanced economies, particularly in technology and professional services, during the same period.

Productivity gains depend on deeper adoption

AMRO estimated that AI could raise Singapore’s potential output through three main channels.

The first is higher total factor productivity as AI automates tasks where adoption is economically viable and reduces labor costs.

The second is additional physical capital accumulation as investment responds to higher output. The third is a potential improvement in effective labor input as AI-assisted learning helps less experienced workers acquire skills more quickly.

Most of the estimated long-term gain comes from the productivity and capital channel, contributing 0.98 percentage points to potential output. The human-capital learning channel contributes another 0.08 percentage points under the baseline scenario.

The estimated long-term gain ranges from 0.47 percent to 2.10 percent depending on the productivity assumptions used.

The timing of adoption also matters. AMRO considered fast, baseline and slow adoption scenarios, with diffusion completed by 2030, 2035 and 2040 respectively.

While the eventual gains are relatively similar, faster adoption would bring the productivity benefits forward. The long-run potential-output gains are estimated at 1.13 percent under the fast scenario, 1.06 percent under the baseline and 1.01% under the slow scenario.

Under the baseline, AI could increase annual potential growth by up to 0.08 percentage point during the adoption phase, particularly between 2029 and 2032.

AMRO cautioned that the estimates represent a one-off comparative-static assessment based on AI capabilities available under the analysis and do not capture additional gains from future advances in AI.

Policy focus on integration and skills

AMRO said Singapore’s policy priority should be to turn broad AI adoption into sustained productivity gains while managing workforce adjustment.

The report identified three areas of focus: deeper integration of AI into firms’ core processes, complementary investment in digital and physical capital, and support for workers affected by changing labor demand.

Government support could place greater emphasis on sector-specific applications, implementation assistance for SMEs, data readiness, governance and organizational change, rather than simply measuring the number of firms adopting AI.

AMRO also recommended greater emphasis on complementary investment in computing capacity, software, data systems, cybersecurity, cloud services and managerial capabilities.

For workers, training and job-redesign support could focus more closely on novice and entry-level employees in occupations where AI is likely to complement rather than replace human tasks.

Existing initiatives such as SkillsFuture AI courses, the SkillsFuture Workforce Development Grant and the National AI Impact Program could help accelerate this transition, AMRO said.

Career conversion, bridging and retrenchment support could also be targeted at workers in highly AI-exposed occupations and recent graduates if entry-level job opportunities weaken.

The report said the aggregate employment impact of AI is expected to be small and temporary, but adjustment costs could still be significant for affected workers.

Ultimately, AMRO said the depth of AI adoption will matter more than its headline adoption rate.

With only 3.8 percent of firms having integrated AI into core business processes, the ability of companies to redesign workflows and develop complementary capabilities will determine how much of the technology’s potential productivity gains Singapore can capture.

AMRO said government support should therefore focus on identifiable market gaps, including access to computing resources, data readiness, SME implementation capacity, workforce training and trusted AI deployment, with progress assessed through measurable outcomes such as deeper adoption, productivity gains and worker transition outcomes.

AMRO flags resource strains, limited spillovers from Malaysia data centers