Global Dialogue on AI Governance

Artificial Intelligence is rapidly reshaping economies, societies and daily life, and while its opportunities are real, so are its risks. From 6-7 July, Geneva welcomed around 4000 stakeholders from more than 170 countries, including member states, technical experts, academics, civil society members and tech companies to hold open, transparent and inclusive discussions on the governance of AI at the Global Dialogue on AI Governance. The Dialogue’s central aim was to ensure that governance does not only reflect the priorities of the most technologically advanced nations, but of all nations, and that the benefits of AI are shared by all and not just a few. 

Over the course of two days, through plenary sessions and side events, and across four thematic clusters, the participants worked around the central question, posed by Secretary-General of the United Nations, António Guterres: will we govern the way that AI will inevitably transform our world, or will we let it govern us? The recently published “Preliminary Report of the Independent International Scientific Panel on AI" – served as a basis and independent evidence base for these discussions and offered a balanced analysis of both risks and opportunities of AI to shape future policy. 

UN Secretary-General António Guterres emphasized the significance of the moment and outlined three warnings for the future of humanity. 

• First, the speed at which AI is advancing is unprecedented with previous technological revolutions. 
• Second, the power over the technology is currently concentrated in a handful of companies and countries, risking hardwiring existing power imbalances and biases into the technology itself and therefore deepening inequality. 
• Third, truth itself becomes distorted, as machine-enabled lies come to be believed as reality. 

While delivering these warnings, Guterres also underscored the potential of the technology, which could become the great equalizer of the 21^st century by compressing decades of progress and development into just a few years. This potential could help deliver the Sustainable Development Goals (SDGs): AI can save lives, warn people of floods and other environmental catastrophes, stop disease outbreaks, support farmers in maximizing their harvest, and provide education for children during crisis. 

A double-edged technology

This dichotomy – between AI’s potential to act as a game changer, boosting development and transforming lives for the better, and its potential to threaten our societies, our striving for equality, and the health of our planet – was woven closely into the two-day dialogue that followed, like a red thread. Certain concepts were central and underscored by stakeholders from all sides of the table, such as the significance of transparency, accountability, safety and agency in AI governance. Amidst concentration among a few companies, apparent knowledge gaps regarding impacts among particularly local stakeholders, the dialogue attempted to begin conversations and to level the playing field to aid the governance process. 

Through its inclusive formats as well as the diversity of actors and initiatives presented, the dialogue created momentum despite the challenges associated with this fast-accelerating technology. 

Human Rights and AI

One of the thematic clusters covered on the second day of the dialogue was concerned with human rights and AI. A takeaway from the work of the Independent International Scientific Panel on AI was that up until now, AI has caused more human rights violations than it has delivered human rights benefits. Areas of special concern are the exploitation of vulnerable populations -  children’s rights and women’s rights –,  the right to non-discrimination and the right to a healthy environment. Especially the environmental footprint of AI, from the location of its data centers, to energy, water, health and ultimately climate, is a key area of concern. While AI benefits can potentially be global, the impacts of the technology are local. Experts emphasized how relevant transparency is in this, and that people who use this technology should ask what the AI they are using actually costs, especially to the local communities. 

On a more positive note, human rights experts highlight that the area of AI and human rights is not uncharted waters. The rights and concepts relevant to it already exist (e.g. Right to non-discrimination, privacy, health, etc.). They need to be applied and integrated, to clarify how they remain relevant in the era of AI. 

Similarly, an expert-level exchange hosted by the ICRC and the Geneva Centre for Security Policy discussed the potential of a Global Compact on Extreme AI risks. At the center was the launch of a study comparing, the U.S., China, the European Union, India, Brazil, South Africa, South Korea and the UAE, that mapped current efforts on AI risk mitigation and areas of convergence. 

These include among others metrics of frontier models, pre-deployment evaluation, rigorous safety testing, frontier weight protection, and incident reporting. At the same time, the report and diverse panel of experts also highlighted the power of particularly countries on the ‘demand-side’ in terms of leverage on safeguards for labor, markets, energy, procurement systems and components. Finally, event drew from successful arms control campaigns such as the Landmine Ban Convention to illustrate the how multi-stakeholder coalitions between small and medium states as well as civil society actors have come together under increasing pressures to contain extreme risks in the past.

Secretary-General António Guterres referred to AI as an experiment on our own societies, without a plan, without our consent. Both the potential and the peril of Artificial Intelligence are too significant and too consequential to be left in the hands of a few, which is exactly why the Global Dialogue on AI Governance took place – and the discussions do not end there. A second session will follow in May 2027, taking place in New York. And in the meantime, these discussions, along with the shared evidence base provided through the report, will offer member states and policy makers a guide as they navigate these increasingly urgent decisions. Parting words that we were left with at the end of the opening were that we will build a future of AI by humanity, with humanity, and for humanity. 

Related Links: 

More information on the Global Dialogue on AI Governance

Preliminary Report of the Independent International Scientific Panel on AI: Evidence-based assessment of opportunities, risks and impacts of AI