by Gergely Rideg
Will we remember this moment decades from now as a paradigm shift in the regulation of artificial intelligence (AI) systems? AI systems and related technology are now as ubiquitous and commonplace as Bluetooth technology was when it was built into a wide variety of devices.
Nowadays, we also see robotic vacuum cleaners[1] with AI, software that analyses the mole[2] , self-driving cars and so on.
Why does it have so much significance if the UN General Assembly adopts a resolution on trustworthy AI?
One could say that the milieu around AI is quite intense and a very exciting topic from the regulators' point of view. In this area, humanity has taken a number of major steps in recent years. On 13 March 2024, the European Parliament approved legislation on AI, which will help the technological innovation of AI while building safeguards to protect our security and fundamental rights. The European legislator has, in our view, taken a giant step towards taming the technological monster that is now a daily topic of debate around the world.[3] It is also true that, there are fewer and fewer countries in the world that do not have an AI strategy at some level.
At this point “the UN General Assembly on Thursday, 21th March 2024 adopted a landmark resolution on the promotion of “safe, secure and trustworthy” artificial intelligence (AI) systems that will also benefit sustainable development for all.”[4]
The importance and news value of the document is naturally enhanced by the fact that the General Assembly is the main policy-making organ of the United Nations. Comprising all Member States, it provides a unique forum for multilateral discussion of the full spectrum of international issues covered by the Charter of the United Nations.
The very fact that such a global organisation is involved in the regulation of AI reinforces the international character and international importance of the problem of the development of AI systems. It also seems clear that the AI phenomenon presents challenges at a global level and positives at a global level.
This occasion may also be of great importance because, so far, the international community in general has not reached a common ground on the adequacy of existing international regulatory solutions for the development of AI systems. The most intense debates and negotiations have been on banning autonomous weapons systems.[5]
Of course, the United Nations itself and its agencies have been working on AI related documents before, but this is the first time that a resolution has been adopted on the subject.
A few examples are worth mentioning here. The UN itself, for example in its 2020 publication, has addressed guidelines that contain substantive concepts both globally and specific to Arab countries. It drew attention to the importance of standards, it spoke about the ethical dimensions of AI, the relationship between the public sector and the private sector in relation to AI systems. This report also included messages highlighting the need for countries to update their education systems to include programming skills, which will be essential in the future. It also calls for the development of skills that cannot be replaced by machines. These include critical thinking, cooperation and social skills, the report says. [6]
In a paper published in April 2022, the organisation addressed the use of AI and automation in contracting. This paper analyses the practical implementation principles of AI systems in contracting. It examines the principles of operation of automated systems in comparison to machine learning AI systems, with a focus on the new legal framework that may be needed for contracts concluded using such techniques. [7]
In addition to the above, the UN has of course also addressed AI in a number of other documents. A report[8] from the year 2022 describes 281 projects[9] on the application of AI in the public interest, taking into account the Sustainable Development Goals. As AI technology is closely related to other information technologies, it is also worth noting that the institution has published resolutions in this field, such as Resolution 55/63 on combating the use of information technology for criminal purposes.
One of the significance of the document now published is that it was adopted without a vote. This means that there was full consensus among Member States on the text. At international level, this confirms the legitimacy and effectiveness of the decision. Also without debate, the UN General Assembly adopted Resolution 70/1 on 25 September 2015, entitled Transforming our world: the 2030 Agenda for Sustainable Development. This resolution established the overarching framework of 17 Sustainable Development Goals (SDGs). The Resolution under discussion pointed out that AI systems will accelerate and enable the achievement of these 17 SDGs.
In the preamble of the resolution, the document reaffirms its link with previous UN documents. Among other things, it states: “Taking note of the efforts of the International Telecommunication Union, in partnership with 40 United Nations bodies, to convene the Artificial Intelligence for Good platform, including its annual summit and the launch of the International Telecommunication Union’s Artificial Intelligence Repository to identify responsible and practical applications of artificial intelligence to advance the Sustainable Development Goals”.
In discussing the importance of the role of AI related to the international law and human rights in maintaining global peace and security, it is crucial to highlight foundational principles recognized by the international community. As stated in the document we analyse:
“Reaffirming international law, in particular the Charter of the United Nations, and recalling the Universal Declaration of Human Rights […]”
The references in the preamble broaden the content of the decision by making it clear that the consensus text is to be read in conjunction with and complementary to the other documents. Thus, the decision not only provides confirmation in itself, but also declares that the results achieved in the past can be linked to the regulatory direction set by the organisation.
When the document also mentions that “the same rights that people have offline must also be protected online, including throughout the life cycle of artificial intelligence systems;” it expands the message a little. It says that the existing, well-established principles of the organisation must be respected in the same way in the online use of AI.
This makes a serious set of rules immediately applicable to AI systems. We believe it is forward-looking to emphasize that protection must apply throughout the entire lifecycle of AI systems. This includes their development, updates, and changes.
In presenting the document, US Ambassador and Permanent Representative to the UN, Linda Thomas-Greenfield said that she hopes the document would not only be a milestone in itself, but that the constructive dialogue on the resolution would serve as a model for other discussions on similar issues in the future.
It is no small achievement that the document includes the recognition that achieving global consensus on technology as soon as possible is a critical task. It also recognised that there are significant differences between countries in tracking the explosion of technology. There is a need to reduce disparities in conditions, opportunities and capacities between developing and developed countries.
Of course, the document also mentions that AI is closely linked to data management and stresses its importance.
It also encourages the private sector to develop practices that respect international law and state law in the design, development and use of AI systems, which are, of course, consistent with the UN Guiding Principles on the Conduct of Business.
Without examining each paragraph of the document, in our view, the landmark nature of the UN resolution can be seen in the following. By repeatedly pointing out the human rights and social coexistence fundamentals that are enshrined in many previous documents, such as the prohibition of discrimination, the resolution places the emphasis on the substance of the human being beyond competition and profit.
As for how AI systems work, we can find out from daily press news and articles in scientific journals how the technology can be used to help us do our jobs better and more efficiently. We also find out how businesses can optimise their operations and how they can increase profits by using certain technologies.
However, to anyone who reads this resolution, the role of technology in sustainable development and in the development of the human subject will become clear. Through the rules, whether prohibitive or restrictive, the essential cornerstones around which the human community has chosen to organise its life in the long term are made clear and unambiguous.
[1] The hidden AI tech that’s packed into robot vacuums - https://www.sciencefocus.com/science/navigation-in-robot-vacuum-cleaners
[2] AI Dermatologist - https://ai-derm.com/
[3] Artificial Intelligence Act: MEPs adopt landmark law - https://www.europarl.europa.eu/news/en/press-room/20240308IPR19015/artificial-intelligence-act-meps-adopt-landmark-law (downloaded: 16.03.2024)
[4] https://news.un.org/en/story/2024/03/1147831
[5] Deeks, Ashley. ”Introduction to the Symposium: How Will Artificial Intelligence Affect International Law?” (2020): 138-140.
[6] Developing an artificial intelligence strategy National guide, United Nations publication issued by ESCWA, United Nations House, Riad El Solh Square, P.O. Box: 11-8575, Beirut, Lebanon. www.unescwa.org
[7] The use of artificial intelligence and automation in contracting, United Nations Commission on, International Trade Law, Working Group IV (Electronic Commerce), Sixty-third session, New York, 4–8 April 2022
[8] https://www.itu.int/dms_pub/itu-s/opb/gen/S-GEN-UNACT-2022-PDF-E.pdf
[9] For example: ”Detecting Fall armyworm through user submitted photos (FAMEWS)”
