Fairwork, Oxford Internet Institute, University of Oxford

I believe that it would be successful if it moves beyond broad principles and produces concrete, actionable outcomes on specific AI governance challenges. While many international discussions on AI already exist, their impact is often limited because they remain too general or aspirational. This dialogue would stand out if participants agree on practical next steps, clear responsibilities, and mechanisms for continued cooperation.

• Protection and promotion of human rights
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches
• Transparency, accountability, and human oversight

I selected these priorities because they directly relate to Fairwork's work on digital labour platforms, AI supply chains, and the conditions of workers who power AI systems. The social, economic, ethical, cultural, linguistic, and technical implications of AI are central to our work because AI systems rely on large networks of data workers involved in tasks such as annotation, content moderation, and data enrichment, often under conditions with limited protections and visibility. I selected interoperability of governance approaches because AI supply chains operate across borders, making greater coordination between governance frameworks important for ensuring accountability and fair labour standards throughout global digital production networks. The protection and promotion of human rights is a priority because workers in AI supply chains are vulnerable to risks including low pay, insecure work, discrimination, and limited representation. Human rights and fair labour standards should therefore be embedded throughout the AI lifecycle. Finally, I selected transparency, accountability, and human oversight because many forms of AI labour remain hidden from public view, while workers are increasingly managed through automated systems. Greater transparency and accountability are necessary to ensure responsible governance and meaningful oversight across AI supply chains.

The listed themes capture many important areas of AI governance. However, because the themes are broad, there is a risk that discussions may not sufficiently address the labour behind AI systems and the conditions of the workers who make these systems possible. AI governance discussions often focus on models, infrastructure, safety, and regulation, while giving less attention to workers involved in data annotation, content moderation, RLHF, and other forms of data work essential to AI development. These workers are frequently part of complex global AI supply chains with limited transparency around labour conditions and accountability. As reflected in Fairwork's AI principles, more explicit attention is needed on issues such as fair pay, fair conditions, fair contracts, fair management, and worker representation.

In the AI supply chain and digital labour sector, governance gaps are contributing to weak protections and limited accountability for data workers. Many governments/organisations/companies still do not have national AI policies or regulatory frameworks that address these issues. Challenges: 1. AI governance discussions often focus on models, infrastructure, and innovation, while the labour underpinning these systems remains insufficiently addressed. As a result, issues such as fair pay, fair conditions, fair contracts, fair management, and worker representation are frequently missing from governance approaches despite being minimum standards as reflected in Fairwork's AI principles. (See Fairwork AI Principles: https://fair.work/en/fw/principles/ai-principles/) 2. The fragmented and cross-border structure of AI supply chains also creates governance challenges. Labour is often outsourced through multiple platforms, vendors, and subcontractors, making responsibility for working conditions unclear and limiting transparency and accountability across the sector. Opportunities: 1. As many governments are still developing national AI strategies and governance frameworks, there is an opportunity to integrate labour standards and supply chain accountability before existing gaps become more deeply embedded. 2. There is also growing interest among governments and international organisations in practical and evidence-based governance approaches. As a not-for-profit research organisation, Fairwork is already working with several governments/organisations/companies on these issues. 3. A key opportunity now is to scale these efforts across more jurisdictions and governance processes through the support of international bodies such as yours.

1. Bring together specialists from different areas of AI governance to hold dedicated discussions on issues such as digital labour and AI supply chains. 2. Create direct exchanges between governments already working on these issues and those still developing national AI governance frameworks. 3. Share practical approaches and lessons drawn from existing frameworks and initiatives such as Fairwork. 4. Bring platforms, vendors, researchers, worker representatives, and policymakers together to address fragmented accountability across AI supply chains. 5. Support common minimum standards across jurisdictions, including fair pay, fair conditions, fair contracts, fair management, and worker representation. 6. Support coordination on cross-border AI labour issues that individual national policies cannot address alone. 7. Encourage governments/companies/organisations to integrate labour and supply chain accountability into national AI governance frameworks earlier in the policy process.

Existing initiative: Fairwork, a joint initiative between the University of Oxford and the Berlin Social Science Center (WZB) is an action research not-for-profit that focuses on improving working conditions in the digital economy, particularly the AI supply chain. Over the last 7 years, Fairwork has supported about 27 million people through 881 company ratings across 42 countries and 443 pro-worker company policy changes. Added value: 1. A key added value of the Dialogue would be its ability to connect labour and AI supply chain issues more directly with broader AI governance discussions, where these issues are often underrepresented. 2. The Dialogue could also encourage companies, governments, and international organisations to give greater attention to working conditions within AI supply chains and create stronger incentives to improve standards and accountability for workers supporting AI systems.

Data annotators, content moderators, RLHF workers, and other workers involved in AI data work remain underrepresented in global AI governance discussions despite being essential to the development of AI systems. Women workers and workers from the Global South are particularly underrepresented, despite making up a significant share of the AI data workforce.

Fairwork's AI principles provide one example of a practical approach to AI governance focused on workers in AI supply chains. The principles establish simple global baseline standards on fair pay, fair conditions, fair contracts, fair management, and worker representation. The approach focuses on measurable minimum thresholds and practical accountability rather than high-level principles alone. It also demonstrates how auditing, certification, benchmarking, and ongoing engagement with companies and organisations can support continuous improvements in labour practices and accountability across AI supply chains. Fairwork's broader work in the digital economy also highlights the value of evidence-based governance approaches that combine worker engagement, public accountability, and cross-jurisdictional benchmarking to improve labour standards in fragmented and cross-border digital labour markets.