AI and green energy security
Description
This dataset supports the study “Can Artificial Intelligence Help Deliver Sustainable Development Goal 7? Evidence on Green Energy Security.” It is a balanced country–year panel of 40 countries from 2000 to 2022 (920 observations). Research hypotheses. The study tests whether national AI capacity improves green energy security (GES) and four related claims: H1: AI has a positive effect on GES. H2: The effect operates through four channels: green innovations (environment-related patents), green energy innovations (renewable energy patents), and public RD&D investment in renewables and in hydrogen and fuel cells. H3: Human capital strengthens the effect. H4: The effect increases with the existing level of GES. Data contents. GES index: A composite of seven physical energy indicators grouped into four pillars—availability, accessibility, affordability and acceptability. Indicators are min–max normalised, with adverse indicators inverted; pillars are equally weighted. A PCA-weighted version is included only as a robustness check. AI capacity: Base-10 logarithm of annual AI publications. Mediators: Environment-related patents, renewable energy patents, and public RD&D in renewables and in hydrogen and fuel cells. Moderator: Human capital. Controls: [list controls, e.g. GDP per capita, trade openness, urbanisation, energy prices]. Key findings. AI raises GES across specifications: 0.034 (fixed effects), 0.053 (IV), 0.072 (CCE-MG), and 0.122 (CS-ARDL long run). The short-run effect is slightly negative (−0.027), so gains appear mainly in the long run. Mediation shares are 86.2% for renewable RD&D investment, 71.1% for green energy innovations, 66.4% for green innovations, and 14.2% for hydrogen and fuel-cell investment. Human capital amplifies the effect: the marginal effect of AI rises from 0.014 to 0.028. The effect is larger where GES is already high: from 0.022 at the 10th percentile to 0.046 at the 90th. Interpretation and use. GES ranges from 0 to 1, with higher values indicating stronger green energy security. AI coefficients represent changes in the index for a one-unit rise in log10 AI publications; for example, 0.1 means a 10-fold increase in publications is associated with a 0.1 higher index score. The data are suitable for panel, heterogeneous-panel and mediation analysis and for benchmarking countries against SDG 7 targets. Patent and RD&D series may contain gaps; check the codebook notes before use. Associations are robust to cross-sectional dependence and endogeneity but should not be interpreted as proof of causation.
Files
Institutions
- North China University of Water Resources and Electric PowerHenan, Zhengzhou