Human Capital, AI, and Labor Commoditization

Auyon Siddiq and Niuniu Zhang
UCLA Anderson School of Management

AI may standardize what workers can produce. Does this make employers view labor as more of a commodity?

What is labor commoditization?

Evidence: AI can compress output differences by raising lower-skilled workers' performance more than higher-skilled workers' performance (Noy and Zhang 2023; Dell'Acqua et al. 2023; Brynjolfsson et al. 2025).

Theory: If AI standardizes worker output, employers may view workers with different skill levels as more substitutable, turning labor into more of a commodity (Fukui 2026).

Implication: Labor commoditization suggests employers might value a worker's human capital (skills, education, employment history) less and become more price sensitive.

This work: AI-based labor commoditization has been theorized but not documented in a real labor market setting; we do so here.

What do we observe?

49,610 worker profiles
2.26M contracts on a large online labor market
2021-2026 study period, includes release of ChatGPT

Worker profile data

We use a language model to read each worker profile as four profile blocks that employers see before hiring.

Self-presentation
Profile title, overview text, skill tags, and stated areas of expertise.
Credentials
Formal education, employment history, certifications, and listed work experience.
Reputation
Job success, completed contracts, ratings, prior employer feedback, and platform history.
Price
The posted hourly rate shown alongside the worker's profile.

Workers' profile information is represented as embeddings.

Each point is a worker's self-presentation embedding. Nearby points have similar profile language. Select a job category to visualize its workers.

We attribute labor demand to human capital and price using machine learning.

ŷiq = ȳq + φselfiq + φcrediq + φrepiq + φpriceiq
i = worker
q = quarter
Average demand The baseline level in a quarter.
Self-presentation How much profile text and skills add to demand.
Credentials How much education and work history add to demand.
Reputation How much ratings and prior feedback add to demand.
Price How much posted hourly rate adds or subtracts.

Demand for workers falls more in high AI-exposure job categories.

After ChatGPT, demand is 7.0% lower in more AI-exposed1 job categories, and 9.6% lower by 2026.

1AI exposure is based on the occupational exposure measure in Eloundou et al. (2024).

How does AI change how employers value human capital vs. price?

Human capital importance falls 7.8% after ChatGPT, while price importance rises 1.1%. By 2026, human capital importance falls 10.1% and price importance rises 1.8%.

The demand premium associated with high human capital shrinks. Demand shifts more toward lower-priced workers.
Raw demand levels for high and low human capital and price groups

Does the change in the demand gap also depend on AI exposure?

Demand shifts are consistent with labor commoditization.

Demand premium for high-human capital workers decreases more for workers in high AI-exposure categories. Demand re-allocation toward lower-priced workers is more pronounced for workers in high AI-exposure categories.
Demand-premium compression by AI exposure

AI-based labor commoditization impacts online labor markets, worker incentives, and worker welfare.

Platforms: Search and ranking may need to adapt if employers view workers as more substitutable.

Workers: Weaker returns to credentials, reputation, and self-presentation may change incentives to invest.

Prices: If price competition intensifies, commoditization may affect worker earnings and welfare.

Traditional employment: Commoditization may also change how employers interpret resumes, experience, and pay expectations.

Link to paper
1 / 11