AI use among German firms surged from 2023 to 2024, particularly in manufacturing and services, led by larger firms and risk-tolerant managers; adopters report expecting meaningful long-term productivity gains.
AI adoption among German firms
AI adoption among German firms increased markedly between 2023 and 2024, concentrated in manufacturing and services and skewed toward larger firms and those led by managers with higher risk tolerance, with adopters reporting expectations of higher long-term productivity.
This paper examines the adoption of Artificial Intelligence (AI) among German firms, leveraging firm-level data from the ifo Business Survey. We analyze the diffusion of AI across sectors and firm sizes, showing a significant increase in AI usage from 2023 to 2024, particularly in manufacturing and services. The survey data allows us to explore not only sectoral patterns of adoption but also the drivers and barriers that firms face, including firm-specific characteristics and industry dynamics. Additionally, we investigate the role of managerial traits, such as risk tolerance and patience, in shaping AI adoption decisions. Finally, we assess the potential productivity impacts of AI at the firm level, with a focus on the expected long-term benefits of AI for different sectors of the German economy. Our findings contribute to the growing body of research on AI adoption by providing new evidence from a non-US context, offering valuable insights for both academia and politics.
Summary
Main Finding
AI adoption among German firms rose markedly between 2023 and 2024, with notable diffusion in manufacturing and services. Adoption patterns vary by sector and firm characteristics, and managerial traits (risk tolerance, patience) help explain which firms adopt. Firms expect positive long‑term productivity benefits from AI, though the size and distribution of those gains differ across industries.
Key Points
- Adoption trend: Substantial increase in reported AI usage from 2023 to 2024, concentrated especially in manufacturing and services.
- Heterogeneity: Diffusion differs across sectors and firm sizes (adoption patterns are not uniform); firm‑level characteristics and industry dynamics are important correlates.
- Managerial characteristics: Managerial risk tolerance and patience are significant predictors of AI adoption decisions.
- Drivers and barriers: Firms report a mix of drivers (expected productivity gains, competitiveness) and barriers (costs, skills shortages, uncertainty) that shape adoption.
- Expected productivity: Firms anticipate long‑term productivity benefits from AI, with heterogeneity by sector—some sectors expect larger gains than others.
- Contextual contribution: Provides non‑US, country‑level evidence (Germany), enriching the literature that is currently US‑centric.
Data & Methods
- Data source: Firm‑level survey data from the ifo Business Survey (modules covering AI usage, barriers/drivers, managerial traits, expectations).
- Coverage: Cross‑sectional comparisons for 2023 and 2024, with sectoral and firm‑size breakdowns.
- Analyses (as reported):
- Descriptive statistics documenting diffusion patterns across sectors and sizes and changes over time.
- Multivariate empirical models relating AI adoption to firm characteristics, industry controls, and managerial trait measures to identify correlates of adoption.
- Assessment of expected productivity impacts using firms’ reported expectations and sectoral comparisons.
- Caveats in inference: Results are based on survey responses (self‑reported adoption and expected impacts); establishing causal productivity effects likely requires longitudinal administrative data or quasi‑experimental variation.
Implications for AI Economics
- Distributional effects matter: Heterogeneous adoption across sectors and firm sizes implies uneven productivity gains and potential shifts in industry competitiveness—aggregated GDP effects will depend on where adoption concentrates.
- Role of management and preferences: Managerial risk attitudes and patience shape adoption decisions, suggesting that models of technology diffusion should incorporate behavioral and preference heterogeneity (not just financial constraints or firm capabilities).
- Policy design:
- Targeted support (training, finance, adoption subsidies) may be needed for lagging sectors and smaller firms to avoid widening gaps.
- Policies that reduce uncertainty (standards, piloting programs, public‑private demonstration projects) and lower the cost of experimentation could increase socially beneficial adoption.
- Complementary investments in skills and organizational change are important—expectations of productivity gains require the right complementarities to be realized.
- Research agenda:
- Move from intentions/expectations to realized impacts using panel or administrative data and causal identification (e.g., instrumental variables, policy shocks).
- Explore complementarities between AI adoption and labor skills, capital investment, and organizational practices.
- Investigate long‑run macroeconomic implications of heterogeneous diffusion in non‑US contexts to inform cross‑country modelling of AI’s economic effects.
Limitations to keep in mind: reliance on self‑reported survey measures (adoption and expected impacts), potential sample selection, and a short observation window for evaluating realized productivity changes.
Assessment
Paper Typecorrelational
Evidence Strengthmedium — The paper provides high-quality, timely descriptive evidence on AI diffusion in a nationally representative firm survey (ifo), and uses multivariate controls to document robust associations; however, it lacks exogenous variation or quasi-experimental identification, relies on self-reported adoption and expectations, and does not cleanly establish causal effects of AI on realized productivity.
Methods Rigormedium — Methods appear appropriate for observational survey work (panel/descriptive statistics, regressions with controls and likely fixed effects); the inclusion of managerial trait measures is a strength, but potential biases from self-reporting, omitted variables, reverse causality, and short panel span reduce rigor for causal claims.
SampleFirm-level responses from the ifo Business Survey covering German firms across manufacturing and services (and other sectors) in 2023 and 2024, including firm characteristics (size, sector), self-reported AI adoption/use, perceived barriers/drivers, managerial trait items (e.g., risk tolerance, patience), and firm-level expectations about future productivity; exact sample size not specified in the summary.
Themesadoption productivity org_design innovation
IdentificationAssociational analysis using firm-level ifo Business Survey data (2023–2024): descriptive diffusion patterns, multivariate regressions controlling for firm size, sector, and other observables, and associations between managerial traits and reported AI adoption; productivity impacts assessed via cross-sectional correlations and firm expectations rather than exogenous variation or experimental causal identification.
GeneralizabilityLimited to German firms — findings may not generalize to other countries with different labor markets, regulation, or digital infrastructure, Short time window (2023–2024) — captures early diffusion but not long-run adoption or dynamic impacts, Survey-based self-reports — adoption, managerial traits, and expected productivity may be measured with error or bias, Potential non-response or selection bias in the ifo survey sample (e.g., underrepresentation of very small firms or certain industries), Correlational design — results do not necessarily generalize to causal effects of AI on productivity across contexts
Claims (7)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| There was a significant increase in AI usage among German firms from 2023 to 2024. Adoption Rate | positive | high | AI usage / AI adoption rate (reported by firms) |
0.3
|
| The increase in AI usage from 2023 to 2024 was particularly pronounced in manufacturing and services sectors. Adoption Rate | positive | high | AI usage / AI adoption rate by sector |
0.3
|
| AI adoption/diffusion varies across firm sizes. Adoption Rate | mixed | high | AI adoption rate by firm size category |
0.3
|
| Drivers and barriers to AI adoption include firm-specific characteristics and industry dynamics. Adoption Rate | mixed | high | AI adoption decision / reported barriers and drivers |
0.3
|
| Managerial traits, such as risk tolerance and patience, play a role in shaping firms' AI adoption decisions. Adoption Rate | mixed | high | AI adoption decision (association with managerial traits) |
0.3
|
| AI is expected to have positive long-term productivity impacts for different sectors of the German economy. Firm Productivity | positive | high | Expected firm-level productivity / anticipated long-term productivity benefits |
0.05
|
| This paper provides new evidence on AI adoption from a non-US context by leveraging German firm-level data (ifo Business Survey). Other | positive | high | Empirical evidence on AI adoption in Germany (contribution to literature) |
0.3
|