The Swiss delegation to the OSCE, in cooperation with the Netherlands, organised a side event on the opportunities and risks of greater involvement of artificial intelligence in cybersecurity, thus actively contributing to the further development of confidence- and security-building measures in the cyber area. This event offered the participating states a first insight into the much-discussed and often misunderstood topic of artificial intelligence (AI) on the margins of the OSCE Informal Working Group on Cyber.

The successful side event attracted 40 delegations and facilitated an interactive debate with powerful contributions from the two panelists from Switzerland, Katharina Frey Bossoni from the Swiss FDFA and Mateo Rojas-Carulla, Chief Scientist and co-founder of Lakera, and from the Netherlands, Dr Roy Lindelauf, Professor of Data Science at the Netherlands Defence Academy. Collaboration between academia, governments and the private sector is crucial for the development of effective policies and regulations, and Switzerland promotes such an inclusive working environment in the OSCE.

The platform allowed the experts to discuss the growing importance of AI and its potential threats, such as its possible misuse for misinformation campaigns and cyberattacks, especially on vulnerable systems such as smart grids. Large-scale language models such as ChatGPT are of particular concern due to their versatility and ease of use.  Securing these models is a major challenge as malicious actors can exploit them with readily available data. The discussion emphasised the need for proactive security measures, clear regulations and international cooperation to counter the evolving threats posed by AI.