Krisenmanagement unter Druck - Wenn KI das Spielfeld neu definiert

The ‘Claude Mythos’ moment and why your crisis management handbook is already out of date

In April 2026, the then Fed Chair Jerome Powell and US Treasury Secretary Scott Bessent spoke personally by telephone with the CEOs of the largest banks. The occasion was Anthropic’s new model, Claude Mythos, an AI capable of identifying 30-year-old security vulnerabilities in seconds. Even the developer considers the model so risky that it has restricted access to it (source: Tagesschau). This moment marks a turning point. Today, crises arise more quickly, are more asymmetrical and strike with a force that traditional crisis management manuals are barely equipped to handle. Any bank in 2026 that relies exclusively on tried-and-tested structures risks not managing the next crisis, but being managed by it.

This is precisely where our workshop comes in.

No rehash of classic crisis management approaches. Instead, an honest examination of what crisis management means in a world where AI is simultaneously a catalyst, a tool and a threat. Four modules combine theory, tools and real-world simulation to form a robust framework for action.

Block 1. Fundamentals and Disruption

‘What constitutes a crisis today? The Claude myth.’ How AI tools such as Claude and ChatGPT can create, accelerate and resolve crises. Plus the new typology of crises in banks, ranging from operational disruptions through reputation and liquidity crises to cyber and AI-induced situations.

Block 2. Building a Crisis Management Framework

“From Reacting to Taking Control.” You will develop the architecture of a robust crisis management team and identify the mechanisms that shift it from reaction mode to control mode.

Block 3. Crisis Communication Under Pressure

“Clear messages when the pressure mounts.” A tried-and-tested toolkit for stakeholder identification, channel selection and message architecture. Including the question of where the use of AI in crisis communication becomes a breaking point.

Block 4. Crisis management team exercise in the context of new realities

“Refine, simulate, analyse.” A comprehensive simulation under time pressure, with incomplete information and AI-driven dynamics. Followed by an honest analysis with concrete lessons learnt for your organisation.

Who is it for?

Executives, crisis management team members, risk managers and those responsible for emergency and crisis management in banks and financial services providers. No prior technical knowledge of AI is required. We translate technology into decision-making logic.

Your added value

You will leave the workshop with a robust understanding of modern crises, a validated crisis management team structure, a tried-and-tested communication toolkit, and the experience of having participated in a crisis management team exercise under the conditions of the year 2026. The warnings from Washington to bank CEOs are not an isolated incident, but a signal. Regulatory authorities, boards and stakeholders expect crisis management in banks to reflect the new reality – not just once the first AI-induced reputational damage has occurred.

Speakers: Fanny Kohls and Ergün Canbay, both from WG-DATA GmbH, Berlin

Fanny Kohls is a qualified risk and compliance manager (M.A.) and a visiting lecturer at Deggendorf Institute of Technology. She brings with her many years of experience in the responsible implementation and consultancy of regulatory matters (MaRisk, DORA, KRITIS, etc.) in the areas of business continuity management, crisis management and contract management.

Ergün Canbay is a certified crisis and emergency manager who assists organisations in preparing for and managing crisis situations under intense time pressure and the need to make rapid decisions. In particular, he supports structured crisis team work, clear decision-making processes and the confident handling of complex stressful situations. As a lawyer and qualified risk and compliance manager (M.A.), as well as the in-house compliance and data protection officer at WG-DATA GmbH, he has extensive experience in the regulatory environment (including the GDPR, MaRisk, DORA and AMLR).