Artificial Intelligence (AI) is on the rise in many industries, including the financial industry, with an increasing number of applications. The benefits offered by AI include everything from improved efficiencies via automation of repetitive tasks up to and including the development of brand-new business models. But the human factor cannot be ignored. That was the consensus between participants in the panel discussion titled “Artificial Intelligence: wind in the sails for business models?” The panellists were Mercedes Bunz, Professor of Digital Culture and Society at King’s College, Mark Branson, President of the Federal Financial Supervisory Authority (BaFin), Michael Hanisch, Head of Technology AWS Deutschland, Marion Höllinger, Spokeswoman of the Executive Board (CEO) of Unicredit Bank and Timotheus Höttges, Chief Executive Officer of Deutsche Telekom. Höttges stressed that “AI is a good tool for analysing data at high speeds, and for providing customised solutions, but the client’s wishes and intentions have to remain in focus, which means we have to disclose when we are using AI in order to achieve maximum transparency.” BaFin President Branson added that Explainable AI and the ability to audit results were also supervisory issues. In the end, he said, it’s people who decide which data are useful for training the AI.
Mercedes Bunz was clear that in order for people to trust AI, they will need to have a human point of contact. “It’s always a collaboration between people and a machine.” Spokeswoman of the Executive Board (CEO) of Unicredit Bank Höllinger had a similar opinion, explaining that society and clients had to be on board if a business wanted to utilise all facets of AI. Only once they had bought in, she said, could business cases be developed. Asked what the most important factor is when it comes to AI-based business models, Head of Technology AWS Deutschland Hanisch answered that a solid data strategy is decisive. “Which data can I use for what purpose? Which data are critical and which are not? Has the client agreed to their use?” Only then would it be possible, he added, to move on to the second step and analyse what could be done with the data and which use-cases could found for it. Telekom CEO Höttges looked to the future. “We are still right at the beginning of AI; the next stage will no doubt be frustrating, and we can expect setbacks.” Only after this stage is over will the technology be truly useful, he said.
