Deutsche Bank relies on partnerships in artificial intelligence
ChatGPT has taken artificial intelligence (AI) to a new level and made it mainstream. The chatbot took just two months to reach 100 million users. For comparison, the social media platform TikTok took nine months, and the music streaming service Spotify took 4.5 years to achieve the same milestone. The use cases of AI, especially generative AI, are diverse.
The German financial industry is also actively exploring the possibilities of AI. Sascha Beck from the Stuttgart-based IT service provider GFT Technologies spoke at the Euro Finance Week in Frankfurt about a veritable race among banks to implement the next AI application.
While artificial intelligence is not new to the banking industry, according to Bernd Leukert, Chief Technology, Data, and Innovation Officer of Deutsche Bank, his institution has been working with AI for years. What's new is generative AI – artificial intelligence that can create content in various forms.
Deutsche Bank collaborates with Google and Nvidia
Deutsche Bank is strongly emphasizing technology partnerships, including collaborations with Google and Nvidia. "For Deutsche Bank, the partnership theme was a fundamental change in strategy," said Leukert. Until 2019, the prevailing view within the bank was to develop, build, and control as much as possible. However, in the end, the realization was that they are not an IT company but a bank. "We cannot invest billions in basic technologies and remain competitive at the same time," said the long-time SAP executive, who is also interested in collaborating with fintechs.
Bernd Leukert, Deutsche Bank
We cannot invest billions in basic technologies and remain competitive at the same time.
AI can help banks streamline and make internal processes more efficient. Additionally, it can assist developers and programmers in better understanding, documenting, and integrating old codes into new environments. Above all, AI is a tool for IT specialists, particularly in terms of quality assurance. It is emphasized, especially for beginners, not to blindly rely on the results of AI, as mentioned by Leukert.
Would AI have spared Deutsche Bank the Postbank trouble?
Leukert also sees the potential of AI to improve customer experience. For example, when it comes to handling a large number of customer inquiries. A human always has to decide whether to handle and process the customer inquiry themselves or pass it on. AI can make that decision in real-time. Did Leukert perhaps have this year's experiences in mind in the context of the bumpy Postbank IT integration?
As part of the "Unity" project, Deutsche Bank migrated 50 billion records from over 12 million customers to Deutsche Bank's systems. The technical data transfer went smoothly, as the bank always emphasizes. However, the bank underestimated how many customer inquiries would come in within this context, leading to significant processing backlogs. The Federal Financial Supervisory Authority (BaFin) received so many complaints that BaFin President Mark Branson publicly reprimanded the bank and sent a special representative to oversee it.
Generative AI may not necessarily have predicted from historical data that there would be so many customer inquiries in this project. However, it might have helped the bank in processing them - an interesting but retrospectively theoretical thought experiment that might have spared the bank a lot of trouble.
Trouble may also loom on the regulatory front regarding the AI issue. While there was consensus on the importance and sensibility of regulation at the Euro Finance Week panel, Tobias Czekalla, Germany head of Visa, cautioned that regulation should not stifle the opportunities of AI. This is crucial for the economic location of Germany.
Unified EU Regulation
One of the biggest risks for Leukert is the issue of uncertainty. It is very difficult to make significant investments when the legal framework is not yet established. While the EU has the issue on its agenda and is currently working on the EU AI Act, a final result is not yet available.
Both Leukert and Czekalla emphasized the importance of a unified European solution in AI regulation. They highlighted the need to avoid individual national approaches since artificial intelligence does not respect national borders. A unified EU regulation is also crucial to compete globally with the USA and China.