Artificial Intelligence Adoption, Governance Disclosure and Responsible Banking in Emerging Markets

Artificial intelligence (AI) is spreading through emerging-market banking faster than the institutions meant to govern it, with consequences for responsible and sustainable finance. We hand-code eleven AI adoption and governance items on a graded 0–3 scale from annual reports of 143 listed commercial banks in eleven emerging markets over 2014–2025 (1582 bank-years), linking every positive score to a verbatim quotation, and merge the index with World Bank, governance and dated regulatory data. Adoption disclosure rises from 5.0% to 83.2% of banks, while governance disclosure reaches only 33.6%, and within banks, adoption trends upward 5.6 times faster (about 3–6 times across robustness tests). Governance disclosure accelerates only in 2024 and 2025, before most banking supervisors had issued any AI instrument and at least as strongly where none existed. Within countries, adoption tracks internet penetration and governance tracks the rule of law, though neither is robust to country-specific trends; data protection laws and national AI strategies explain neither series. Disclosed governance shows no association with credit risk or profitability, and the hypothesised link between adoption and capital ratios (H2) is not supported. The results identify a pre-institutional stage in which AI governance is neither demanded nor supplied, and offer a ready disclosure template.