The Magyar Nemzeti Bank (MNB) previously published a comprehensive digital transformation guidance for domestic banks and stipulated that they prepare a dedicated artificial intelligence (AI) sub‑strategy by June 30, 2026. The guidance has been applicable to credit institutions since April 1, 2026.
Scope of the AI sub‑strategy
According to the MNB, the AI sub‑strategy must cover all phases of introducing and operating the technology: from data governance and research and development through risk management and internal controls to customer communication and system decommissioning. The regulator’s primary objective is to ensure transparent, ethical and secure operation.
Banks that already use AI or plan to introduce it must supplement their existing digital transformation strategies with an AI‑focused sub‑strategy. Responsibility for the area and its implementation must be assigned to a leader who has appropriate expertise and decision‑making authority.
Internal rules, lifecycle management and data governance
Banks are required to establish internal regulatory and control frameworks that cover the full lifecycle of AI solutions, from R&D to deployment, operation and decommissioning. Because AI depends on data, the MNB insists on strict data governance, especially for high‑risk applications.
Institutions must ensure training and test datasets are of high quality, accurate, free from bias and legally compliant. Models used in development and procurement must be safeguarded against external manipulation and cyber threats, and the reproducibility of results must be assured.
High‑risk systems and human oversight
The MNB specifically requires continuous human oversight and the possibility of intervention for high‑risk AI systems. Banks must also be prepared to handle technical failures effectively.
The regulator demands that only systems whose identified data protection, legal, financial, model‑ or cyber‑security risks have been reduced to an acceptable level may be deployed in production. Banks are expected to carry out a detailed assessment of their institution‑specific AI risk profile.
Expertise, training and third parties
Responsible AI use presupposes adequate internal expertise. The MNB requires regular, role‑appropriate theoretical and practical training for staff. Clear internal rules must be established for the workplace use of third‑party tools, particularly generative AI models.
Transparency and customer rights
The MNB guidance emphasizes transparent communication inside and outside the organization. Banks must inform customers clearly and understandably when a service involves interaction with AI or when content they see was produced with AI assistance. Customers must be offered the option of traditional, human service, the right to have AI‑driven decisions reviewed, and a dedicated channel to report any errors.
Deadlines
The MNB’s expectations have been in force since April 1, 2026. Credit institutions were required to produce and submit their detailed AI sub‑strategies for the first time by June 30, 2026 — the submission deadline set by the regulator.
Why it matters
With these measures the MNB aims to make the use of AI in banking more transparent, ethical and secure, reducing data protection, legal and cyber risks for customers and the financial system. The guidance provides banks with detailed direction for the responsible introduction and operation of AI technologies.
Tags
data protection, MNB, Magyar Nemzeti Bank, cybersecurity, banks, artificial intelligence, digital transformation, regulation, AI, generative artificial intelligence
