Addressing the safety and security challenges of complex AI systems is critical to fostering trust. It is widely recognised that realising the benefits of the technology requires coupling innovation with stringent safety measures. Following the UK’s and South Korea’s AI Safety Summits and anticipating France’s AI Action Summit in 2025, this session will explore international collaborations and the progress toward frameworks to manage and mitigate AI risks effectively. It will focus on risk assessment and mitigation through robust safety and fairness measures, ensuring that AI development and deployment remain safe, and secure throughout the entire lifecycle of systems. Key topics will include independent evaluations, testing, and red-teaming; the monitoring and measurement of AI risks; assessing GenAI risks such as hallucinations, misuse, lack of traceability, and harmful output; understanding frontier AI risks; such as dual-use applications and the role of various stakeholders in ensuring AI safety.

Possible points to discuss:

• Outcomes from recent international AI summits (Bletchley – Seoul), including The Frontier AI Safety Commitments by industry
• Development of a network of international AI safety institutes and how they will work together and with other relevant stakeholders
• Strategies for monitoring and measuring existing and emerging AI risks including multidisciplinary approaches to defining risk thresholds and model risk assessment
• Technical and procedural guardrails for ensuring the safe development, deployment, and use of AI and addressing risks arising from uses outside of intended purpose, intentional misuse, or unintentional misuse.
• Approaches to maintain robust, secure, and safe AI systems throughout their entire lifecycle and defining delineating roles and responsibilities along the AI value chain: stakeholders trust and empowerment through testing, accountability, transparency
• Role of the scientific community in advancing AI safety and research and boosting international research collaborations for AI risks mitigation
• Role of standardisation in promoting safety measures
• Responsible information sharing
• Policy treatment of open-source vs proprietary AI, and of long-term risks vs present AI harms
• Understanding Frontier AI risks while addressing current risks

The advent of AI is accelerating breakthroughs that promise to catalyse global development and societal change. However, disparities in AI capabilities can reinforce imbalances in wealth and power,  exacerbating the digital divide between developed and developing countries and, indeed, within many countries. A cohesive and inclusive approach will be needed to develop equitable models that effectively address diverse regional, national, and local needs. This panel will explore the potential of trustworthy AI to drive growth and prosperity for individuals and society, or if poorly delivered, to have opposite effects. It will discuss the role of governance and international cooperation in shaping a prosperous and equitable AI future, and how supporting countries in forming homegrown AI ecosystems and regulatory frameworks can advance global development objectives.

Possible points to discuss:

• The importance of more equitable partnerships with developing countries to ensure that AI advances positive change globally
• Strategies to harness AI for global progress and how equitable access and inclusive global AI governance can concretely be enabled
• The extent to which collaboration on R&D projects focused on solving large-scale global problems can stimulate international cooperation and drive AI innovation
• The landmark ‘Verona Process on Digital Development’ where G7 countries committed to work with developing countries to strengthen local AI digital ecosystems.
• The development of equitable models that effectively serve diverse populations
• Bridging the digital divide between and within countries by democratising computing power, developing open and secure data models, nurturing talent and strengthening capacities to advance AI local ecosystems – and the role of the private sector in achieving this
• Ensuring that AI systems reflect countries’ and regions’ cultural and linguistic diversity to prevent bias and discrimination and expand access
• The role of Free Trade Agreements or Digital Economy Agreements to increase access to AI

This session will explore how industry-specific and cross-sector partnerships in AI are fueling innovation, sustainability, and growth across economies, by taking live and planned cases to enrich our understanding of applied AI. Given the level of investments pouring into the field of AI, the session will address the key steps needed to develop an expanded global market for AI tools and applications, with a strong focus on how SMEs and startups can also incorporate AI responsibly.

• The importance of defining clear roles within the AI ecosystem / value chain
• How Enterprise AI differs from Consumer AI and whether these differences are understood and considered by policymakers when looking at regulating the technology
• Overview of the anticipated evolution of AI systems in both the medium and long term and the crucial role of foresight in helping policymakers anticipate and keep pace with technological developments.

As artificial intelligence continues to evolve, combining policymaking with technical expertise, and establishing robust international standards will ensure interoperability, data quality, reliability, measurement, and processes to achieve transparency, safety, security, and the ethical use of AI systems globally. This session will discuss progress made through international efforts to create and implement AI standards, focusing on how these standards can facilitate governance, promote innovation, and protect individual rights. It will explore the gaps, needs, and collaborative opportunities in AI standardisation and how initiatives can be harmonised globally.

Possible points to discuss:

• The state of AI standards development in international standards organisations, including the development of EU harmonised standards to implement the EU AI Act
• Addressing divergence in AI standards and the implications for international governance and trade
• Exploring methods to accelerate the development and implementation of AI standards with international partners.
• The role of standards in bridging differences in approaches to law, regulation, and governance and preventing fragmentation among systems
• The need to strengthen the dialogue between the tech industry, regulators and standardisation bodies to develop practical and applicable standards and the importance of inclusive participation in the standards-making process to ensure broad representation and acceptance.
• Balancing global standards with regional specificities and requirements.
• Integrating technical and broader societal considerations in AI standards development.
• The effect that standards can have in the open-source v closed-source debate
• The roles of the private sector and of governments in developing standards and the importance of research-driven standards

The European Union has taken a pioneering step with the AI Act, the world-first legislation designed to regulate AI based on a risk-based approach, and which seeks to lead the global discourse on trustworthy and ethical AI. As the application of the AI Act is staged over two years, it will take time before it has a meaningful impact. In the meantime, the technology will continue to develop and deploy at speed, with the success of the law dependent on consistent implementation and strong enforcement . Meanwhile, AI development and deployment will be intricately entangled with complex geopolitical dynamics, with global businesses having to navigate different regulatory strategies and bottlenecks in the AI value chain, such as digital infrastructure, computing power, data and talent.

The session will discuss the implications of the AI Act for the AI landscape in Europe and globally, considering the implementation timeline, the rapid pace of AI development and the current global race for technological dominance that influences the direction of AI research, investment, supply chain and regulation. It will also explore the role that the European AI Office and the AI Board will play in ensuring the success of the AI Act, in supporting the European AI ecosystem, as well as in shaping international agreements on AI.

Possible points to discuss:

• A comprehensive understanding of the EU AI Act, implementation timeline implications and challenges, and updates on the AI Pact
• What to expect from the upcoming delegated or implementing acts and guidelines
• The important role of regulatory sandboxes in facilitating the development and testing of innovative AI systems.
• The potential impact of the AI Act on the competitiveness of European AI startups and SMEs
• The recent release of the innovation strategy on Artificial Intelligence to establish ‘AI Factories’
• How the AI Office plans to foster international cooperation and align with global AI regulatory efforts aiming to shape global agreements on AI
• How the AI Office will work with global institutions to monitor the evolving AI ecosystem, technological and market developments, and the emergence of systemic risks.