On March 11, 2024, the General Assembly of the United Nations (UN) published its final declaration „Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development“. While this resolution is non-binding under international law and merely contains guidelines and recommendations, it is nonetheless noteworthy. This is because the General Assembly emphasizes the need to reach a global consensus on safe and trusted artificial intelligence (AI) systems in accordance with international law, in particular the Charter of the United Nations, the Universal Declaration of Human Rights, and the 2030 Agenda for Sustainable Development.
See also: n2406592.pdf (un.org) And General Assembly adopts landmark resolution on artificial intelligence | UN News
“So let us reaffirm that AI will be created and deployed through the lens of humanity and dignity, safety and security, human rights and fundamental freedoms.” This is how Linda Thomas-Greenfield, U.S. Ambassador and Permanent Representative to the United Nations, summed up the draft resolution prior to its adoption. Went to say that the “inclusive and constructive dialogue that led to this resolution would serve as a model for future conversations on AI challenges in other arenas,” and urged UN member states to “to govern this technology rather than let it govern us”.
The Europeans moved faster
However, the UN decision is not the first in the field of AI legislation. One step ahead of the UN legislators are the European Union and other countries. The European Union adopted a proposal for an AI law in December 2023 (see also: First AI law: finally some clarity). In March 2024, the EU finally passed its AI law. It requires developers of AI technologies to disclose their training data and conduct rigorous testing of the technology’s impact and functioning.
Unanimously approved
The United Nations has unanimously adopted the resolution initiated by the United States of America to promote the safe and trustworthy development of artificial intelligence. This shows how AI has “transcended usual geopolitical divisions,” said a senior US government official. The UN resolution is the result of the largest international effort to date to reach agreement on artificial intelligence technology. However, there are no enforcement mechanisms in case countries fail to comply with the resolution.
US officials have been negotiating with over 120 UN member states since fall 2023. Despite “lots of heated conversations” with opponents such as Russia and China, the end result was a text that no one could object to and was therefore adopted by consensus by the UN General Assembly.
Safe and trustworthy AI
The UN resolution on artificial intelligence was adopted by more than 120 countries. The adopted text uses the term „safe, secure and trustworthy artificial intelligence systems“ more than two dozen times. In addition, it is emphasized right at the beginning of the resolution that the field of artificial intelligence is constantly evolving. Therefore, an ongoing discussion on the different needs and capacities of different countries, especially developing countries, is necessary.
Protection of Human Rights and Fundamental Freedoms
The General Assembly begins its resolution by reaffirming international law and the Charter of the United Nations and recalling the Universal Declaration of Human Rights. It notes that, even in the digital age, human rights and privacy must be particularly protected and promoted in the use of digital technologies.
The United Nations emphasizes that human rights and fundamental freedoms must be respected, protected and promoted throughout the life cycle of AI systems. They therefore call on all Member States to use only AI systems that can be operated in accordance with international human rights standards. “The same rights that people have offline must also be protected online”.
In support of Sustainable Development
In addition, the United Nations General Assembly reaffirms its 2030 Agenda for Sustainable Development and recognizes that safe and trustworthy artificial intelligence (AI) systems can contribute to the achievement of all 17 United Nations Sustainable Development Goals (THE 17 GOALS | Sustainable Development (un.org)). Therefore, the General Assembly resolves to promote safe and trustworthy AI systems that contribute to the full implementation of the 2023 Agenda for Sustainable Development and the achievement of all economic, social and environmental Sustainable Development Goals.
See also: AI technologies: Working for sustainability
The General Assembly is also convinced that secure and trustworthy AI systems can be a driver of digital transformation, a force for peace, and a bridge to the digital divide between and within countries. It is crucial, however, that this be done human-centered. Developing countries in particular need technical and financial support to build capacity.
Cooperation with developing nations
The General Assembly calls upon Member States to work with developing countries to improve their digital infrastructure and access to technological innovation, and to strengthen the capacity of the least developed countries in the field of artificial intelligence. In addition, developing countries should be involved in international processes and fora related to the governance of artificial intelligence systems.
Protection against discrimination, disadvantage and abuse
The United Nations seeks to protect people from all forms of discrimination, disadvantage and abuse. It calls on Member States to promote safe and trustworthy AI systems in an inclusive and equitable manner for the benefit of all, leaving no one behind. The aim is to protect privacy and personal data, as well as intellectual property rights, including copyrighted content. At the same time, innovation should be encouraged.
Member States should also create opportunities to bridge the digital gender divide and to integrate equality for people with disabilities, gender equality and equality between people of different races into policy making.
Oversight of data management
The United Nations General Assembly recognizes that data is critical to the development and operation of AI systems. It therefore urges Member States to promote international cooperation, collaboration and support for data management. It emphasizes that UN Member States should strengthen their cooperation to create fair, open, inclusive and non-discriminatory frameworks, and therefore encourages Member States to promote a level playing field for small businesses, entrepreneurs and technical talent.
In addition, the United Nations wishes to promote effective, accessible, adaptable and internationally interoperable technical tools, standards or practices to determine the origin and creation of public content. This should enable users to detect manipulation of information, distinguish authentic digital content from that created or manipulated by artificial intelligence, or determine its origin. In addition, the UN aims to promote the transparency, predictability, reliability and accountability of decisions made by AI systems.
Encouraging the exchange of information
The General Assembly also calls for the promotion and intensification of communication between different stakeholders. According to the UN, the equitable, inclusive, responsible and efficient use of digital public goods is essential to realize the potential of safe, secure and trustworthy AI systems for sustainable development.
Basis of the resolution
The adopted UN resolution builds on an earlier international agreement on AI. However, it was criticized for failing to include developing countries.
In November 2023, the United States, China, the European Union, the United Kingdom, and more than 20 other countries signed the so-called Bletchley Declaration, which aimed to avoid existential security risks of the technology and to promote international cooperation in research.
Artificial Intelligence Summit at Bletchley Park, England
In early November 2023, a summit on the risks of artificial intelligence brought together heads of government from six continents at Bletchley Park, a historic site in northwest London that rose to fame during World War II.
The European Union and 27 countries, including the United States and China, agreed on a landmark declaration in Britain to limit the risks and harness the benefits of artificial intelligence. US Vice President Kamala Harris, UK Prime Minister Rishi Sunak and other world leaders joined technology leaders including Elon Musk, Sam Altman, DeepMind founder Demis Hassabis and Microsoft CEO Brad Smith at the two-day event.
Participants agreed to allow experts from the UK’s new Institute for AI Safety to test new AI models for risks before they are released to the public. Unfortunately, there are few details about how the tests will be conducted. Moreover, the agreements are largely voluntary.
AI resolution for civil use only
AI technology is already being used on battlefields and in election campaigns, and could even influence the outcome of wars. However, the AI resolution explicitly does not apply to artificial intelligence in the military sector, UN diplomats stress.
Another UN General Assembly is scheduled for mid-April to May 10, 2024, which will include the (provisional) agenda item „Lethal Autonomous Weapons Systems“. It is time for the international community to face these challenges.
Please find the German translation of this article here: UN-Resolution zu KI: Menschlichkeit und Würde müssen an erster Stelle stehen