As part of the 2026 CEPR European Workshop on Household Finance, organized jointly with the Bundesbank and hosted at the Bundesbank’s training center in Eltville, a panel discussion was held on “Financial Advice in the Era of Artificial Intelligence.” The session was moderated by Falko Fecht, Head of the Bundesbank’s Research Centre, and featured the following panelists: Prof. Dr. Fritzi Köhler-Geib, Member of the Bundesbank’s Executive Board, Prof. Dr. Christine Laudenbach, Leibniz Institute for Financial Research SAFE, Professor Dr. Michael Haliassos, Institute for Monetary and Financial Stability (IMFS) and CEPR, and Prof. Dr. Peter Oertmann, Ultramarin GmbH and TU München.

A central theme was the increasing integration of AI-powered tools—such as chatbots, robo-advisors, and personalized financial assistants—into mainstream banking and investment services, as well as the direct use of AI by households (eg ChatGPT) for financial advice. These technologies are now capable of analyzing transaction data in real time, providing tailored budgeting and investment advice, simulating financial scenarios, and even dynamically adjusting investment portfolios based on user behavior and preferences. The panel noted that while these innovations have the potential to democratize access to high-quality financial advice, especially for previously underserved groups, they also raise important questions about data privacy, trust, and the risk of amplifying biases or creating new forms of discrimination.

The discussion highlighted both the promise and pitfalls of AI-driven advice. On the positive side, AI can help close the "advice gap" by making financial guidance more accessible and affordable, potentially reducing traditional biases in human advice (such as gender or wealth-based discrimination). However, research presented at the workshop showed that AI systems can themselves inherit or even exacerbate biases present in their training data, and that users may not always understand or trust AI-generated recommendations—especially for complex or high-stakes decisions like mortgages or retirement planning. The importance of maintaining a "human in the loop" was emphasized, with evidence suggesting that hybrid models (combining AI and human advisors) are often preferred and more effective.

From an industry perspective, the asset management sector is seen as being in an "advanced installation phase" of AI adoption, with significant investments in infrastructure and pilot projects but relatively few fundamental changes to core processes so far. However, AI is predicted to transform asset management, as the potential of processing very large datasets, elevating research productivity, creating highly customized investment products and enabling better informed capital market decisions at significantly lower costs is real. 

The panel also discussed the implications for regulation, monetary policy, financial stability, and consumer protection, noting that the scalability of AI could amplify both benefits and risks. Looking ahead, the panelists called for more research on the real-world use of AI in financial decision-making, the dynamics of trust and adoption among different demographic groups, and the development of regulatory frameworks that balance innovation with privacy and fairness.

In summary, the workshop underscored that AI is already reshaping the landscape of household finance, but its full impact will depend on how well the industry, regulators, and researchers address issues of bias, transparency, data security, and the human-AI interface. The future of financial advice is likely to be hybrid, with AI augmenting but not fully replacing human expertise, and ongoing vigilance required to ensure that technological advances translate into genuinely better outcomes for all households.

This is an AI-aided summary.