AI is reshaping Pakistan’s IT jobs: routine coding and documentation are being automated, boosting productivity—particularly for junior staff—while raising demand for AI skills and concern about reduced entry-level opportunities and wage pressure for mundane roles.

Main Finding

AI is reshaping Pakistan’s IT labour market primarily by transforming tasks and job content rather than causing immediate mass layoffs. Among 20 IT professionals interviewed in Islamabad and Rawalpindi, most reported productivity gains from AI (especially for routine coding, documentation, and data processing) alongside rising demand for new skills (AI literacy, prompt engineering, data interaction, oversight). Economic consequences are likely to be heterogeneous: workers with AI skills are expected to capture higher pay, while those in routine roles face wage pressure and reduced entry-level opportunities, raising concerns about growing job insecurity.

Key Points

• Productivity and task change
  • Respondents reported AI increased productivity (automating repetitive coding, documentation, data processing), freeing time for creative/complex work.
  • Common development tasks (programming, testing, maintenance) are being altered: less routine coding, more AI‑validation, system integration, and oversight.
• Skills and training
  • New and growing skills: AI literacy, prompt engineering, data interpretation, critical evaluation of AI outputs, and continuous upskilling.
  • Need for employer-led and lifelong learning; concerns about limited training availability in Pakistan.
• Wages and distributional effects
  • Expectation of greater wage dispersion: premium for AI‑capable/high‑skill workers; downward pressure on pay for repetitive/routine roles.
  • Productivity gains do not automatically translate into higher wages for all; employers may raise expectations without proportional pay increases.
• Job security and labour market entry
  • Interviewees did not anticipate a sudden wave of layoffs but feared fewer entry‑level positions and rising precarity.
  • Overall view: AI is more likely to change jobs (task composition, quality) than to destroy them outright.
• Policy and organizational recommendations (as reported)
  • Regular training/upskilling programs, stronger university–industry collaboration, and ethical AI use standards to manage the transition.

Data & Methods

• Design: Qualitative, interview‑based study.
• Sample: 20 IT professionals working in Islamabad and Rawalpindi, Pakistan.
• Instrument: Work‑related open‑ended interview questions covering five topics — productivity, changes in jobs/roles, skills and training needs, pay/compensation, and job security.
• Analysis: Thematic synthesis of respondents’ experiences and perceptions (no large‑scale quantitative measurement reported).
• Contextualization: Findings are discussed against task‑based technological change literature (e.g., Acemoglu & Restrepo), productivity studies (Brynjolfsson et al., Gambacorta et al.), and reports from OECD/ILO.

Limitations (inferred from method) - Small, non‑representative sample limits generalizability to Pakistan’s broader IT sector. - Self‑reported perceptions—no direct measurement of wage changes, employment flows, or productivity at firm level. - Cross‑sectional qualitative design cannot capture dynamic/longitudinal effects.

Implications for AI Economics

• Supports the task‑based framework: AI reallocates tasks within jobs, yielding reinstatement (new tasks) and displacement (automatable tasks) effects; net employment and wage outcomes depend on their relative strength.
• Heterogeneous productivity gains imply heterogeneous returns to labor:
  • Expect skill‑biased effects and increased wage dispersion within the IT sector.
  • Junior/less‑experienced workers may see large productivity boosts from AI tools, but this could compress entry pathways if fewer novices are needed.
• Policy levers and research priorities:
  • Workforce policy: prioritize scalable upskilling/reskilling programs, employer‑funded training, and university–industry partnerships to prevent skill bottlenecks and mitigate inequality.
  • Labour market monitoring: collect firm‑level and administrative data on hiring, wages, and entry‑level vacancies to detect shifts in demand for junior roles and routine tasks.
  • Redistribution and safety nets: consider targeted support for displaced or downgraded routine workers if wage pressures materialize.
• Empirical research needs:
  • Larger quantitative studies in Pakistan to measure AI’s causal impact on wages, employment composition, and entry‑level hiring.
  • Longitudinal designs to track career trajectories of workers who adopt AI skills versus those who do not.
  • Employer‑side research to understand how firms capture productivity gains and translate (or not) into hiring and compensation decisions.

Overall, the study offers ground‑level evidence that AI is transforming IT work in Pakistan through task reallocation and skill complementarities, reinforcing the need for policy responses focused on training, labour‑market monitoring, and equitable distribution of AI gains.