news

recently, the united nations released an important report on the governance of artificial intelligence (ai), calling on the international community to work together to establish a truly global ai supervision and governance mechanism.

the report, written by the un secretary-general’s high level advisory body on ai, sets out a series of recommendations aimed at addressing the risks posed by ai while ensuring its transformative potential is equitably shared globally.

the report emphasizes that the need for global governance of ai is unquestionable. the raw materials for ai technology, including key minerals and training data, come from all over the world.

general ai technology will be deployed across borders, generating multiple applications around the world. at the same time, the rapid development of ai has concentrated power and wealth around the world, bringinggeopoliticsand economic impact.

more importantly, no one currently fully understands all of ai’s inner workings, nor can they fully control its output or predict its evolution.

at the same time, policymakers are not held accountable for developing, deploying, or using systems they do not understand. these decisions are likely to have global negative consequences and downstream impacts.

to address existing gaps in ai governance arrangements, the advisory body makes several key recommendations:

• establish an independent international ai science group composed of experts from different disciplines, serving on a voluntary basis in their personal capacity.

• launch a biennial intergovernmental and multi-stakeholder policy dialogue on ai governance, which could take place on the margins of existing un meetings, providing a platform for the un’s 193 member states to discuss risks and build consensus for action.

• establish an “ai standards group” to develop and maintain a manual of definitions and applicable standards for evaluating ai systems.

• the establishment of a global ai fund, expected to be managed by a new independent governance structure, aims to bridge the gap in ai development, especially supporting development in countries in the global south.

• develop a global ai data framework and establish “market-shaping data management and exchange mechanisms” to promote the thriving development of local ai ecosystems around the world.

• establish an office on ai within the un secretariat to better unify these initiatives in a lightweight, flexible structure.

the report also recommends that the un focus on poorer countries, particularly those in the global south, to enable them to benefit from ai and participate in its governance.

these measures include establishing ai standards and data-sharing systems, and allocating training and education resources to help countries with ai governance.

some of the recommendations could be implemented through the existing global digital compact, which aims to address digital and data divides between countries.

wendy hall, emeritus professor of computer science at the university of southampton, the only british expert on the advisory body, stressed: “ai must develop for the benefit of humanity, and without governance it has the potential to evolve in ways that are harmful to society.

we urgently need a global approach to the challenges of ai to ensure it benefits everyone.”

however, ai governance is not currently a truly global conversation. the report points out that while hundreds of guidelines, frameworks and principles on ai governance have been adopted by governments, companies, alliances and international organizations, none of them has a truly global impact. this leads to problems with representation, coordination and enforcement.

this unequal distribution of power is also reflected in the entire international system.

un secretary-general antónio guterres noted at the world economic forum in davos in january 2024 that many member states were still under colonial rule when the un was founded and therefore have “very little voice” in today’s discussions.

he called for reform of the international system established after world war ii to enable effective action on global challenges.

in march 2024, the united states submitted a resolution to the united nations calling on member states to embrace the development of "safe, secure, and trustworthy ai."

in july of that year, china proposed its own resolution, emphasizing cooperation in ai development and making the technology widely available. all un member states signed both agreements.

differences in views on ai between wealthy nations have already led to cracks in the market. the european union has introduced sweeping ai regulations, and its data-use controls have prompted some u.s. companies to limit the availability of their products.

the us government’s laissez-faire approach has also led the state of california to propose its own, stricter ai regulations.

in the face of these challenges, the report of the un advisory body on ai paves the way for building a common position.

chris russell, a professor at oxford university, said: “anchoring the analysis in human rights is very persuasive.

it provides a broad and powerfulinternational lawfoundation and focus on the specific harms people suffer.”

however, ai is developing at such a rapid pace that the un will not be able to manage global cooperation alone.

“the un clearly has an important role to play in ai governance, but it needs to be a distributed structure,” he said. “countries also need to be directly involved.”

finally, while governments may see ai as a way to gain strategic advantage, many scientists agree on their concerns about ai.

earlier this week, a group of prominent western and chinese scholars jointly called for greater cooperation on ai safety issues following an ai safety conference in vienna, austria.

overall, this un report provides a blueprint for international cooperation on ai governance.

however, as alondra nelson, a member of the advisory body, said: "the devil is in the details."

in the future, how the united nations and its member states choose to implement this blueprint for cooperation will determine the success or failure of global ai governance.

references:

https://arstechnica.com/ai/2024/09/united-nations-wants-to-treat-ai-with-same-urgency-as-climate-change/

https://www.un.org/techenvoy/ai-advisory-body

https://www.computerweekly.com/news/366611263/UN-body-urges-globally-inclusive-and-distributed-AI-governance

operation/layout: he chenlong