Generative AI could reshape many Philippine service jobs, but exposure does not equal inevitable job loss: most exposed roles show complementarities and early adoption has led to task reconfiguration and skills shortages rather than mass layoffs. Without timely skills, governance and social-protection policies, however, the country risks polarization or disruptive displacement by 2025–2035.

Main Finding

AI’s labor-market impact in the Philippines is not preordained by technology alone. While a large share of employment—especially in service sectors and BPO-related occupations—is exposed to generative AI, most of those exposed jobs show high complementarity with AI (i.e., scope for augmentation). Current adoption has been cautious, producing task reconfiguration and emerging skills gaps rather than mass displacement. Outcomes over 2025–2035 will hinge on policy choices: with supportive skills, governance, social protection, and innovation policies the country can realize inclusive upgrading; absent such policies, risks include polarization or disruptive displacement.

Key Points

• Evidence triangulation: the paper combines three complementary streams (task-level capability assessment, occupational exposure/complementarity using Philippine labor-force data, and firm-/worker-level adoption evidence) to avoid deterministic forecasts.
• Exposure vs. complementarity: many jobs are automatable in part (high exposure), but most of those roles also display high complementarity—tasks where human judgment, supervision, or domain knowledge remain crucial—pointing to augmentation rather than immediate job loss for most workers.
• Sectoral concentration: exposure is concentrated in service sectors and occupations tied to the BPO industry, making these sectors central to the policy response.
• Adoption patterns so far: firms have adopted generative AI cautiously; impacts observed include task reconfiguration, shifts in work processes, and skills gaps (not large-scale layoffs).
• Scenario framework (2025–2035): three policy-contingent futures are plausible:
  • Augmented Services Economy — inclusive productivity gains, broad upskilling, and improved job quality.
  • Dual-Speed Labor Market — uneven adoption and adjustment, with polarization between technologically complemented workers and those left behind.
  • Disruptive Automation Shock — rapid displacement in vulnerable occupations, significant social strain, and weak labor-market absorption.
• Policy window: there is a limited opportunity to steer adoption toward inclusive outcomes through timely interventions.

Data & Methods

• Triangulated evidence base:
• Task-level analysis of current generative AI capabilities to identify which types of tasks AI can already perform in practice (e.g., text generation, summarization, coding assistance, translation).
• Occupational exposure and complementarity analysis using Philippine labor-force microdata to estimate what share of employment is exposed and to assess whether exposed occupations are likely to be substituted or complemented by AI.
• Firm- and worker-level evidence (surveys, case studies, adoption indicators) to observe real-world adoption patterns, reconfiguration of tasks, and emerging skills gaps.
• Integrated framework: links AI technical capabilities to occupational task bundles and situates these links within institutional mediators (training systems, labor regulations, firm behavior, social protection).
• Scenario development: a prognostic, policy-contingent scenario approach (not deterministic forecasting) that maps plausible trajectories under different policy and institutional responses for 2025–2035.
• Empirical emphasis: concentrates analysis on sectoral and occupational heterogeneity, and on the complementarities/substitutability of tasks rather than assuming full job automation.

Implications for AI Economics

• Complementarity matters as much as exposure. Economic models and policy analysis should distinguish task-level substitution from augmentation; aggregate exposure rates overstate likely short-term job loss when complementarity is high.
• Institutions shape outcomes. Labor-market institutions, training systems, firm hiring practices, and social protection determine whether AI leads to upgrading, polarization, or displacement — emphasizing the need to model policy endogeneity.
• Sectoral and distributional focus required. The concentration of exposure in services and BPO implies potentially large sector-specific impacts and distributional consequences (wage pressure, job quality changes) that aggregate analyses can miss.
• Policy levers can change trajectories. Investments in targeted skills development, active labor-market policies, innovation support for firms, and adaptive social protection can materially shift outcomes toward the Augmented Services Economy scenario.
• Measurement and research priorities:
  • Move beyond static automation scores to dynamic, task-level exposure measures that account for complementarity and evolving AI capabilities.
  • Collect longitudinal firm–worker matched data to observe task reallocation, wage effects, and heterogeneous adoption pathways.
  • Monitor adoption indicators (firm uptake, task reconfiguration), labor-market signals (vacancies, reemployment rates, wage dispersion), and training outcomes to evaluate policy effectiveness.
  • Study informal-sector dynamics and regional heterogeneity in adjustment capacity.
• Policy timing is crucial. The Philippines faces a narrow window to implement complementary policies that can steer AI adoption toward inclusive upgrading rather than disruptive adjustment.