Key facts
- Microsoft's chief responsible AI officer Natasha Crampton stated that 'AI sovereignty' means integrating local values and cultures into AI systems, not operating in isolation.
- Crampton highlighted the widening digital divide in AI adoption between developed and developing nations.
- She pointed to initiatives like the Lingua project in Africa, a partnership involving Microsoft, the Gates Foundation, and Google.org, to collect local language data for AI models.
- Microsoft is adopting a 'community-first' approach to building AI infrastructure, managing resource consumption to avoid impacting local electricity rates or water supplies.
- Crampton emphasized the need for international collaboration in AI safety and testing, citing Europe's AI Office as a positive example.
Microsoft's chief responsible AI officer, Natasha Crampton, has articulated a nuanced definition of 'AI sovereignty,' asserting that it does not equate to technological isolation but rather the integration of local cultures, values, and norms into AI systems while leveraging global advancements.
Speaking at the UN AI for Good Summit, Crampton expressed concern over the widening digital divide in AI adoption between the Global North and South, warning against an exacerbating 'AI divide.' She emphasized that true AI sovereignty empowers communities with the autonomy and skills to control their AI-driven futures.
To bridge this gap, Crampton highlighted initiatives like the Lingua project, which has expanded to Africa in partnership with the Gates Foundation and Google.org. This project focuses on collecting local-language data to enable foundational AI models to understand idiomatic phrasing and cultural nuances.
Crampton also underscored the importance of collaboration between the private sector, governments, and international bodies like the UN to ensure AI is developed and governed safely and equitably. She pointed to the UN's Global Dialogue on AI Governance as a mechanism to reflect the priorities of all nations.
Microsoft is also promoting a 'community-first' approach to building AI infrastructure, aiming to be a 'good neighbor' by managing resource consumption to prevent impacts on local electricity rates and water supplies. The company is also involved in the 'digital emblem' initiative with the ICRC and the International Telecommunications Union, designed to protect critical infrastructure from cyberattacks.
Regarding Europe, Crampton commended the European AI Office's efforts in international coordination on AI safety and testing. She urged regulators worldwide to maintain flexibility, acknowledging that AI regulations must adapt to the rapidly evolving technology and its associated risks.