Governance and development of AI: regulatory requirements and good practices

AI governance and oversight – Model selection – Performance evaluation – In-house model deployment – AI data management – Managing AI risks

The Regulatory Framework for AI Deployment in Switzerland and the European Union

• Impact of the EU AI Act in Switzerland
• How to concretely implement the principles of fair, ethical, accountable, and transparent AI use
• Drafting an AI policy: what is the purpose?

AI Governance and Oversight: Roadmap and Observed Practices

• Update existing risk management policies and procedures to strengthen AI governance
• Review governance structures, roles and responsibilities, policies and processes to address AI-related risks
• Risks of outsourced AI

AI principles in practice

• The principles of artificial intelligence.
• Model selection tailored to the task in hand.
• Measuring the performance of an AI model.
• In-house model deployment with Retrieval-Augmented Generation.
• Case study: Anti-money laundering with neural networks.

Data & AI: What you must know about Data for AI

• The universe of data
• Ethical and social dimensions
• Why quality matters
• AI data management

AI Risk Management: Specific Challenges and Approaches to Managing AI Risks

• AI risk factors: financial risk, operational risk, regulatory risk, reputational risk
• AI risks vs. generative AI risks (e.g., hallucination risk)
• Inventory of AI systems in use: scope definition challenges
• Identify AI use and related risks throughout the bank
• Assess the materiality of AI-related risks: impact on the bank, clients, and stakeholders