The Swiss Financial Market Supervisory Authority (FINMA) has published the findings of a recent survey on artificial intelligence (AI) use among financial institutions in Switzerland. The survey was conducted between late November 2024 and mid-January 2025.

FINMA surveyed approximately 400 licensed institutions. These included banks, securities firms, insurance companies, insurance intermediaries, fund managers, and market infrastructure providers.

Widespread Use of AI

Around half of the surveyed institutions currently use AI or have pilot applications underway. An additional 25% plan to adopt AI within the next three years. On average, each respondent uses five AI applications and has nine more in development. Larger institutions tend to have more AI use cases than smaller ones.

Among AI users, 91% make use of generative AI tools, such as modern chatbots. Many institutions rely on services from large technology providers. FINMA noted a growing dependence on external vendors, especially among smaller firms that often do not develop AI tools in-house.

An #AI Future For Financial Sector#MachineLearning #DeepLearning #BigData #Fintech #ML #DL #Robotics #insurtech https://t.co/OkT8ssYKVz pic.twitter.com/NbqUn8Rvk2

— AI (@DeepLearn007) September 20, 2017

This reliance increases outsourcing risks. FINMA highlighted these risks in its 2024 Risk Monitor, stating that outsourcing critical functions to third parties remains a major operational concern for the financial sector.

You may find it interesting at FinanceMagnates.com: AI Will Take Your Job — CEOs Say Deal with It — What Can You Do?

AI Governance and Risk Management

About 50% of institutions have established a formal AI strategy. Most organisations manage AI under existing governance systems, focusing on data protection, cybersecurity Cybersecurity Cybersecurity is a blanket term that refers to the protection of computer systems and networks from the theft.More broadly speaking, cybersecurity can also represent countermeasures against damage to hardware, software, or electronic data, as well as from the disruption or misdirection of the services they provide.It was not long ago that the term cybersecurity not exist as it was first used in 1989. In today’s vernacular cybersecurity, refers to measures taken to protect a computer or computer Cybersecurity is a blanket term that refers to the protection of computer systems and networks from the theft.More broadly speaking, cybersecurity can also represent countermeasures against damage to hardware, software, or electronic data, as well as from the disruption or misdirection of the services they provide.It was not long ago that the term cybersecurity not exist as it was first used in 1989. In today’s vernacular cybersecurity, refers to measures taken to protect a computer or computer Read this Term, data management, and enterprise risk management Risk Management One of the most common terms utilized by brokers, risk management refers to the practice of identifying potential risks in advance. Most commonly, this also involves the analysis of risk and the undertaking of precautionary steps to both mitigate and prevent for such risk.Such efforts are essential for brokers and venues in the finance industry, given the potential for fallout in the face of unforeseen events or crises. Given a more tightly regulated environment across nearly every asset class, One of the most common terms utilized by brokers, risk management refers to the practice of identifying potential risks in advance. Most commonly, this also involves the analysis of risk and the undertaking of precautionary steps to both mitigate and prevent for such risk.Such efforts are essential for brokers and venues in the finance industry, given the potential for fallout in the face of unforeseen events or crises. Given a more tightly regulated environment across nearly every asset class, Read this Term.

Institutions identified key risks linked to AI. These include concerns about data quality, data protection, and the explainability of AI decisions. Respondents also raised issues about data security, incorrect outputs, and outsourcing-related risks.

Sector Breakdown

Of the 187 institutions currently using AI, 100 are banks and securities firms. Seventy-five are insurance companies and intermediaries. The remaining 12 include fund managers and financial market infrastructure operators.

“FINMA will ensure that the use of new and innovative technologies on the Swiss financial market is in line with the regulatory framework through transparent and technology-neutral authorisation and supervisory activities,” the regulator stated.

“In doing so, FINMA will follow the principle of ‘same business, same risks, same rules’. The objective of the survey was to track the development of the use along with the governance and risk management of AI in order to ensure a risk-based supervisory approach.”