Sofia Romansky is a Strategic Analyst at the Hague Centre for Strategic Studies in the Hague and Master's graduate of International Relations and Diplomacy with a focus on social and internet technologies, the role of information in conflict, and the military and societal applications of artificial intelligence. She will focus on some case studies and international governence as she is also the Project Coordinator for the Global Commission on Responsible AI in the Military Domain. The integration of artificial intelligence (AI) into the military domain is no longer a matter of science-fiction, it is already actively a part of our reality and is (re)shaping the character of war. In light of this realisation, this presentation aims to lay out why AI is garnering such attention, the opportunities that AI presents for the military domain, but also the challenges created both by ethical dilemmas and obstacles for adequate governance. Concluding with a review of ongoing international governance initiatives and several key next steps forward, this presentation suggests that there is still time, but also an urgent need, to create responsible guidelines for AI in the military domain which maximises the benefits while mitigating risks, in accordance with core ethical and legal values.

Cyberspace has many unique opportunities and challenges for all actors that exist in it, this especially includes states and governmental organisations. From ensuring good communications and preventing attacks on critical infrastructure, to intelligence gathering and the disruption of rival actors' capabilities, states have many goals and roles to play within the digital domain. In this lecture, we will explore the roles, goals, and methods of state actors in the international cyber realm, diving into the many challenges that state actors face in this ever-evolving environment, as well as their geo-political motives for engaging in offensive cyber activities. Tim Zonsveld is an Information Security Advisor for the municipality of The Hague. In his daily work, he ensures that the municipality remains digitally secure. He supports the city's Chief Information Security Officer (CISO) and helps develop and refine information security policy at a municipal level. In 2006-2007, he also served as Secretary and Vice-President on our association's Board.

Our digital society relies on software, and in the past decades, we have produced huge amounts of code. This production speed does not intend to stop, and in recent years software developers have gained new and shiny productivity enhancers: generative AI tools, e.g., ChatGPT from OpenAI or Copilot from GitHub. Trained on the huge amounts of code from the past, these tools spit out new programs and functions at unprecedented speed and help developers to code much faster. Yet, this is also in the caveat: being trained on the huge amounts of code from the past, generative AI tools also perpetuate historical security weaknesses from these codebases. These weaknesses, called vulnerabilities, can be misused by the attackers and enable new devastating cyber incidents. In her talk, Olga Gadyatskaya will start by introducing the basic idea of what a software vulnerability is. She will then discuss recent research studies that evaluated the security of the generated code, demonstrating the widespread security issues reproduced by generative AI tools. She will also discuss why vulnerabilities get perpetuated from training, and will conclude with some advice on how to ensure code security when using generative AI tools.

In his talk, Valerio Viscardi explores the complexities and potential of information sharing within the EU’s cybersecurity framework. He begins by highlighting the critical role of threat intelligence sharing in strengthening collective cyber resilience. Key topics include the benefits of collaboration, current challenges such as legal fragmentation and trust issues, and the significant impact of regulations like the NIS2 Directive and GDPR on data-sharing practices. He underscores how legislative constraints affect both public and private sectors, limiting the effectiveness of threat intelligence exchange.