Artificial intelligence (AI) has become increasingly vital across various sectors, including finance, with the Swiss financial market witnessing a significant surge in its importance. As per the strategic objectives for 2021-2024, FINMA, Switzerland's financial regulatory authority, embraces innovation while closely monitoring associated risks.

Despite AI's transformative potential, most financial institutions are still in the experimental phase, while some have adopted advanced applications necessitating robust risk management processes. The emergence of AI-driven solutions like ChatGPT has further fueled interest in AI within the financial sector, hinting at forthcoming changes.

However, FINMA identifies four key areas posing challenges in AI adoption within finance:

1. Governance and Responsibility: With AI potentially making autonomous decisions, ensuring accountability becomes complex. Transparency issues arise, especially in complex, company-wide processes, like those involving ChatGPT, where factual correctness may be difficult to ascertain. Thus, clear roles, responsibilities, and risk management processes are imperative.
   
2. Robustness and Reliability: AI's reliance on vast datasets raises concerns regarding data quality and model accuracy. Instances like the failure of AI algorithms in predicting Covid-19 underscore the importance of ensuring the accuracy and reliability of AI models, especially when deploying them autonomously.
   ‍‍
3. Transparency and Explicability: The complexity of AI models often obscures the reasoning behind decisions, posing challenges for verification and explanation. Institutions must prioritize making AI applications transparent and explainable, aligning with customer expectations and regulatory requirements.
   ‍‍
4. Non-Discrimination: AI's utilization of personal data for risk assessment or customizing services can inadvertently lead to biased outcomes if certain groups are underrepresented in the data. Addressing such biases is crucial to avoid unintentional discrimination and mitigate legal and reputational risks.

FINMA collaborates with stakeholders and academia to develop expectations and guidelines for AI applications in finance. It emphasizes continuous monitoring and dialogue to stay abreast of AI developments and ensure adherence to regulatory standards.

In conclusion, while AI presents promising opportunities for innovation in the Swiss financial market, navigating its complexities requires careful consideration of governance, reliability, transparency, and fairness to realize its full potential while mitigating associated risks.

Source: FINMA Risk Monitor 2023