Not yet human-reviewed
This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.
Understanding Firms’ AI Efforts and Their Economic Impact
Citation. Babina, T. (2026). Understanding Firms’ AI Efforts and Their Economic Impact (NBER Working Paper No. 35123). National Bureau of Economic Research. https://doi.org/10.3386/w35123
Summary. This working-paper review develops a framework for selecting and interpreting firm-level measures of AI effort, distinguishing invention, adoption, internal capability building, outsourcing, actual use, and investor perceptions. It surveys evidence on how these measures relate to firm growth, valuation, productivity, risk, labor, competition, and financial-market outcomes. The review argues that divergent findings often reflect different measurement targets rather than direct contradiction.
Key claims
- Firm-level AI datasets measure different economic objects, so measure selection determines which questions about AI adoption and effects can be credibly answered.
- AI adoption is uneven, with larger firms and firms with more high-skilled workers often adopting more aggressively; reported investment associations with sales growth and valuation commonly emerge with lags.
- Aggregate firm-level productivity effects are harder to detect than task-level effects, consistent with adjustment costs and organizational complements, while task substitution can coexist with positive net firm-level employment growth.
Evidence & limitations
- Narrative literature review and measurement framework drawing on firm-level surveys, labor measures, patents, disclosures, telemetry, deals, and infrastructure indicators; it does not contribute a new causal estimate.
- NBER working paper circulated for discussion and comment and not peer-reviewed; associations in the reviewed literature do not by themselves establish causal effects.