A model warns that consumer demand switching to AI services can widen wage gaps in non-traded sectors, but Indian transport data show no clear post-2007 spike in inequality — uptake of AI ride services stayed below 0.7%, muting measurable effects.

Artificial Intelligence, Demand Switching and Sectoral Wage Gap

Shreya Roy, Bibek Ray Chaudhuri · Fetched March 30, 2026 · Economic & Political Weekly

semantic_scholar theoretical low evidence 7/10 relevance DOI Source

The paper shows theoretically that consumer switching to AI-enabled non-traded services can increase wage inequality among otherwise homogeneous workers, but Indian transportation-sector data around a 2007 structural break do not show a significant rise in inequality, likely because AI service uptake remained very low.

Artificial intelligence (AI) induced services are a reality in India and other developing countries. It is an opportune time to assess the impact of such technologies on wage inequality. While doing so, we must distinguish labourers who are complementary to AI and from those who are substitutes. In this paper, we develop a demand-switching model which brings in an additional factor. Proportion of consumers, who easily adapt AI-induced services, will influence the pricing of such services. This would further impact the wages across traditional and non-traditional services. The primary aim of this research is to examine how demand switch towards AI-enabled goods and services influence wage differentials among homogeneous workers in the non-traded sector of an open economy. Especially we discuss the case of the transportation sector. Employing a Finite Change General Equilibrium model, we introduce AI as a technological shock in the non-traded sector. Modelling its effects through price adjustments. Our findings suggest that a shift in preference towards non-traded AI services exacerbates income inequality among previously homogeneous workers. Furthermore, empirical validation of our theoretical framework employs an endogenous structural break analysis, identifying 2007 as the break year for AI introduction in India, coinciding with the emergence of AI-powered services such as Windows Live and AI ride-hailing apps like Ola and Uber. Our investigation into wage disparities within the transportation sector, measured by the Gini coefficient, shows that despite AI’s introduction, inequality has not significantly worsened. This can be attributed to sluggish demand switching from non-AI to AI-based services in India. Data reveals that less than 0.7% of the Indian population uses AI-induced ride services, underscoring the gradual nature of AI’s societal integration and subsequent impact on economic variables.

Summary

Main Finding

A demand-switching model and finite-change general equilibrium (GE) analysis show that a consumer-driven shift toward AI-enabled non-traded services can increase wage inequality among previously homogeneous workers in the non-traded sector. Empirical analysis for India, however, finds that after an identified structural break in 2007 the transportation sector’s wage inequality (Gini) has not significantly worsened — attributed to very slow consumer switching to AI services (estimated AI-ride usage < 0.7%).

Key Points

Distinction between workers complementary to AI and those who are substitutes is crucial for predicting wage outcomes.
The paper introduces a demand-switching channel: the proportion of consumers who adopt AI-induced services affects pricing of those services, and thus relative wages across traditional and AI-enabled (non-traded) services.
Theoretical result: when consumer preferences shift toward AI-enabled non-traded services, price adjustments can raise wage dispersion among homogeneous workers in that sector.
Empirical claim: an endogenous structural break analysis identifies 2007 as the start of measurable AI introduction in India (authors note contemporaneous emergence of services like Windows Live and AI-enabled ride-hailing such as Ola/Uber).
Measured outcome in the transportation sector: Gini-based wage inequality did not significantly increase post-break, consistent with low adoption—less than 0.7% of the population using AI-induced ride services.
Conclusion: potential for AI to amplify inequality exists, but realized effects depend critically on consumer adoption speed.

Data & Methods

Theoretical model: a finite-change general equilibrium framework for an open economy with a non-traded sector. AI is modeled as a technological shock operating through price changes in the non-traded sector; demand switching (consumer adoption) is an explicit parameter.
Key theoretical mechanism: interaction of technology shock + heterogeneous consumer adoption → price adjustments → differential wage effects for workers who are complements vs substitutes to AI.
Empirical strategy:
- Endogenous structural break analysis to identify the timing of AI introduction in India (break year estimated as 2007).
- Sectoral case study: transportation sector analyzed for wage inequality dynamics.
- Inequality measurement: Gini coefficient of wages within the transportation sector over time.
- Use of adoption/usage data to estimate penetration of AI-enabled ride services (reported <0.7%).
Interpretation: theoretical model predicts widening inequality under sufficient demand switching; empirical data do not show such widening because adoption has been too slow to generate the full equilibrium effects.

Implications for AI Economics

Demand/adoption is a first-order determinant of AI’s labor-market impacts: even if technology enables substitution or complementarity, consumer uptake rates govern price evolution and wage effects in non-traded services.
Sectoral heterogeneity matters: non-traded sectors (e.g., local services, transportation) can experience large relative-wage effects if AI adoption becomes widespread; early-adopter sectors should be monitored for rapid distributional change.
Policy relevance:
- Policies that accelerate or slow consumer adoption (e.g., regulation of platforms, subsidies, digital access programs) can indirectly shape wage inequality outcomes.
- Targeted labor policies (retraining, social insurance) should anticipate distributional outcomes conditional on plausible adoption trajectories.
Research recommendations:
- Collect better, higher-frequency measures of consumer adoption of AI-enabled services to forecast distributional impacts more precisely.
- Extend the model to dynamic adoption paths and spillovers across traded/non-traded sectors.
- Disaggregate labor by tasks/skills (not just homogeneous workers) to capture more realistic complement/substitute relationships.
Empirical caution: the absence of large inequality changes in early data does not rule out future effects if adoption accelerates; monitoring and scenario analysis are warranted.

Assessment

Paper Typetheoretical Evidence Strengthlow — The core contribution is a theoretical GE model with plausible mechanisms, but the empirical support is weak: the structural-break approach provides only correlational timing evidence, the 2007 'break' attribution to AI is contestable, the wage-inequality test uses aggregate Gini measures that are noisy and subject to confounders, and uptake of AI services is very small so any measurable effect is likely limited—there is no strong quasi-experimental identification (e.g., instruments, diff-in-diff, or micro-level causal variation). Methods Rigormedium — Theoretical work appears methodologically sound in using a finite-change GE framework to introduce demand switching and price adjustment channels, but empirical methods are modest: the structural-break and aggregate Gini analysis are standard but limited for causal claims, and the paper lacks micro-level identification, robustness checks against alternative break dates, or controls for concurrent structural changes in the economy. SampleAggregate/sectoral data for India focused on the non-traded transportation sector; time-series of wages and computed Gini coefficients for transportation wages around 2007; auxiliary descriptive statistics on AI-ride-hailing uptake (reported as under 0.7% of population). Exact data sources, sample years, and microdata coverage are not fully specified in the abstract. Themesinequality labor_markets adoption IdentificationDevelops a finite-change general equilibrium model in which demand-switching toward AI-enabled non-traded services changes relative prices and hence wages; empirical corroboration attempted via an endogenous structural-break test (identifies 2007 as break) and time-series/aggregate analysis of Gini coefficients for wages in India's transportation sector, supplemented by descriptive uptake statistics (<0.7% using AI ride services). GeneralizabilitySingle-country analysis (India) — results may not apply to high-income or different institutional contexts, Focused on one non-traded sector (transportation), limiting applicability across sectors, Model assumes homogeneous workers and specific demand-switching behavior, which may not hold in practice, Empirical period and low AI adoption rates reduce ability to detect effects — findings may not generalize to later periods with higher uptake, Aggregate Gini results mask within-sector heterogeneity (urban vs rural, formal vs informal workers)

Claims (8)

Claim	Direction	Confidence	Outcome	Details
Artificial intelligence (AI) induced services are a reality in India and other developing countries. Adoption Rate	positive	high	presence/adoption of AI-induced services	0.12
The proportion of consumers who adopt AI-induced services influences the pricing of those services and through price adjustments will further impact wages across traditional and non-traditional services. Wages	mixed	high	wages (across traditional and non-traditional services) and service prices	0.12
A shift in preference towards non-traded AI services exacerbates income inequality among previously homogeneous workers in the non-traded sector (model finding). Inequality	positive	high	income inequality / wage differentials among homogeneous workers	0.12
Endogenous structural break analysis identifies 2007 as the break year for AI introduction in India. Adoption Rate	positive	high	identified structural break year for AI introduction	2007 (break year) 0.12
Despite AI’s introduction, wage inequality in the transportation sector (measured by the Gini coefficient) has not significantly worsened. Inequality	null_result	high	Gini coefficient of wages in the transportation sector	0.12
The lack of a significant worsening in transportation-sector inequality can be attributed to sluggish demand switching from non-AI to AI-based services in India. Adoption Rate	negative	high	rate of demand switching / adoption	0.12
Data reveals that less than 0.7% of the Indian population uses AI-induced ride services. Adoption Rate	negative	high	share of population using AI-induced ride services	less than 0.7% of the Indian population uses AI-induced ride services 0.12
AI’s societal integration in India is gradual, and therefore its impact on economic variables (like wages and inequality) is also gradual. Adoption Rate	null_result	high	pace of AI integration and consequent economic impact	0.12