Innovative business models and the use of new technologies in the financial sector offer promising opportunities, but also risks – for supervised entities and for consumers.

Artificial intelligence and machine learning

Companies in the financial sector are increasingly using artificial intelligence (AI), especially generative AI. This is true of the entire value chain. Many initiatives regarding the use of generative AI are still in the testing or pilot phase.

In actual practice, testing is focused in particular on large language models (LLMs). LLMs are used mainly as assistance systems. Common applications include GPT-based internal chatbots for employees, AI systems for the preparation of documents and AI assistance systems for developers (coding). Generative AI also offers potential in terms of customer contact. To date, however, its use in this area has been relatively limited.

Banks and insurers, on the other hand, are already using more traditional forms of AI and machine learning (ML) to detect money laundering and fraud as well as in back-office processes. One example is “dark processing ”, which is used by insurers in claims processing. In risk management, AI and ML are increasingly being used to process data and validate risk models. In the highly automated securities industry, AI and ML are primarily used to improve processes in trading, advice, risk management and compliance.

Limits and risks in practice

The use of AI also has its risks: AI models are often not explainable and are difficult to verify. There is also a risk that AI models are based on data imbalances or prejudices. This can lead to a distortion of the results (bias) and may have unintended consequences, such as discrimination against customers. In addition to the damage they can cause to customers, discriminatory actions on the part of AI can result in liability and reputational risks for AI system providers and operators.

Risks to financial stability may arise from new dependencies on third parties, in particular on large cloud and AI model providers. The reason for this is a strong market concentration on a small number of providers. Furthermore, AI could give rise to herd behaviour. The risk is that general use of the same AI information processing and automated trading strategies will result in a very large number of market participants behaving in a similar way.

Quantum computing: early protective measures necessary

For there to be a breakthrough in powerful and stable quantum computers, technological hurdles still need to be overcome. There must be sufficient computing power available, for example, and the complex physical framework conditions necessary for state-of-the-art quantum computing (QC) must be met. Common quantum computer concepts require cooling at temperatures close to absolute zero, for example.
Various applications are already being researched in the financial services sector, in areas such as risk management and portfolio optimisation. In future, the industry will be dependent on the services of large technology providers for such applications.

A breakthrough in quantum computing will pose risks for the financial sector. Now is the time for companies in the financial sector to take appropriate IT security measures. One example for the urgency of this situation is the “harvest now, decrypt later” problem.

Cryptoassets: upward trend after scandals

Following scandals and the collapse of major providers in 2023, the crypto market had shrunk to a tenth of its 2021 trading volume. By the end of October 2024, total market capitalisation had doubled compared to the previous year (see Figure 15). This growth was largely due to the rise in the price of Bitcoin, which was spurred by the introduction of spot Bitcoin ETFs in the US in January 2024. Other cryptoassets also achieved significant price gains in the same period.

Following the presidential elections in the US, the crypto market has continued to grow and has since reached new all-time highs, most recently at the beginning of December at approximately 3.6 trillion euros (see Figure 15). Investors’ expectations that the legal framework for crypto trading in the US would be relaxed in future, which could lead to stronger demand, have had a significant impact on this price performance.

Bitcoin remains the dominant cryptoasset. At the end of 2024, the Bitcoin price reached a new record high of just under 103 thousand euros. This constitutes a 2.7-fold increase compared to the previous year. Among the key trends in the crypto market are the growth of liquid staking
With liquid staking, users receive alternative derivative tokens in return for the tokens they have deposited. They can then use these alternative tokens and thus, for example, generate returns on their staked assets. and the increasing tokenisation of both traditional financial instruments and real-world assets (RWAs).

Figure 15: Market capitalisation (all cryptoassets)

Source: BaFin diagram using data from coingecko, as at 1 December 2024

Cryptoassets can influence the traditional financial market if they are used as collateral for loans or if the reserves of stablecoins tie up significant portions of the available short-term money market instruments. Despite the high trading volume of cryptoassets, there was no systemic risk to the German financial market at the end of 2024 (see Figure 16).

Figure 16: Development of trading volumes in cryptoassetsn

Source: BaFin diagram using data from coingecko, as at 1 December 2024

More articles

Risks in BaFin's Focus 2025
Foreword by the President

Main Risks in BaFin’s Focus

1. Risks arising from corrections on the real estate markets
2. Risks arising from significant corrections on the international financial markets
3. Risks arising from corporate loan defaults
4. Risks arising from cyber incidents with serious consequences
5. Risks arising from inadequate money laundering prevention
6. Risks arising from market concentration due to the outsourcing of IT services

Trends

2. Sustainability
3. Geopolitical turmoil