AI is transforming banking, but its rapid adoption raises security and financial-stability concerns. Regulators in the US and Europe are stepping up scrutiny of advanced AI systems, especially those that can make decisions without constant human oversight.

The growing scrutiny reflects fears that cyberattacks, uncontrolled automated systems and dependence on external technology providers could disrupt essential banking services.

Why Regulators Are Concerned

Banks increasingly use artificial intelligence to detect fraud, identify cybersecurity threats and automate complex operations. However, advanced AI agents can perform tasks independently, introducing risks that traditional security systems may struggle to manage.

Financial supervisors are particularly concerned about situations in which automated systems exceed their authority or respond unpredictably to cyber threats.

Recent incidents involving AI-assisted hacking have increased these concerns, while the International Monetary Fund has warned that major cyberattacks could undermine confidence in financial markets.

Regulators are therefore examining whether banks have adequate safeguards before expanding their use of autonomous technology.

American Banks Face New Questions

Three major US financial regulators are investigating how institutions deploy advanced AI models.

The Federal Reserve, Office of the Comptroller of the Currency and Federal Deposit Insurance Corporation are examining the use of AI agents in cybersecurity operations.

Their questions focus on whether automated systems can identify threats, make decisions and respond safely without excessive independence.

Supervisors also want to know whether banks can monitor AI activity, review decisions and immediately disable systems when necessary.

The use of foreign-developed AI models is receiving additional attention.

For now, these inquiries primarily involve gathering information rather than imposing new requirements.

European Authorities Focus On Recovery

European regulators are placing particular emphasis on banks' ability to restore operations after serious cyber incidents.

The Bank of England is encouraging major financial institutions to examine independent recovery systems that can operate separately from their main technology infrastructure. These arrangements could allow essential services to restart even if primary systems become compromised. However, creating and maintaining completely independent recovery infrastructure can be expensive.

The European Central Bank has also instructed lenders to prepare plans addressing AI-related cyber threats by the end of October 2026.

These plans include stronger backup arrangements and procedures for restoring critical operations.

Technology Dependence Creates Additional Risks

Another concern involves banks' increasing reliance on external AI developers and technology providers.

Financial institutions must understand which systems depend on third-party services and whether failures could affect essential operations.

European supervisors are asking banks to map these relationships and demonstrate that AI systems can be safely disconnected and restarted.

Some institutions still have significant weaknesses in these arrangements.

Experts warn that inadequate recovery procedures could leave customers unable to access accounts or complete payments following a major attack.

What This Means For Customers

Stronger oversight could encourage banks to improve cybersecurity, develop more reliable recovery systems and maintain human control over important automated decisions. However, these measures may require substantial investment as financial institutions integrate increasingly sophisticated technology.

For customers, the main concern is ensuring that essential services remain secure and accessible.

As artificial intelligence becomes more deeply embedded in banking, regulators are seeking to ensure that innovation advances alongside effective safeguards against operational failures and cyber threats.