“The field of AI has evolved rapidly, producing an overwhelming amount of information and making it difficult to decipher hype from reality. “This can fuel confusion, prevent common understanding, and benefit large AI companies at the expense of policymakers, civil society, and the public.”
That’s one of the key findings of the United Nations High-Level Advisory Committee on Artificial Intelligence, a group of 39 experts created specifically by the organization to develop guidelines for a global governance strategy for the technology. Its first report was released Thursday, and its main recommendation to the UN is to create an “international science panel” that would increase transparency and reduce the marketing component of AI.
This international group of researchers would aim to disseminate “impartial and reliable scientific knowledge” in order to “remedy information asymmetries between AI laboratories and the rest of the world.”
The group of advisors points out that this lack of reliable data on the real capabilities of this technology, its potential and its risks, is one of the causes of the absence of standards to control it. “Currently, there is no global framework that regulates AI, which implies that its development, in the hands of a few multinationals, can affect the majority of the population without their participation in decision-making,” they point out. The biggest gap is in the countries of the South, where 118 countries have not participated in any “remarkable” initiative to set the rules of AI, they warn.
The Advisory Committee therefore urges the UN to create the first global AI oversight structure, based on international cooperation. The “international scientific group” that helps to better understand these systems would be the first piece of this building. Its missions would be to prepare periodic analyses, but also to monitor the progress of AI with “ad hoc reports on emerging issues, including the emergence of new risks or significant gaps in the governance landscape,” the Advisory Committee requests.
The group would be coordinated by a new UN Secretariat, the report details, which would also be responsible for overseeing implementation of the rest of the recommendations at the global level.
Global policy to control AI
The UN High-Level Advisory Committee on Artificial Intelligence is composed of 39 people from 33 different countries, with experience in academic research on artificial intelligence, but also in the policy and business side of this technology. The co-chair of the committee is Carme Artigas, former Secretary of State for Digitalization and AI in the last legislature.
Its members also include Mira Murati, chief technology officer at OpenAI; or Virginia Dignum, a Portuguese academic who is a reference in AI ethics worldwide. The Committee’s final report appears after a year of work in which 250 written proposals from organizations and experts were analyzed on the current state of this technology and what the architecture should be to govern it at a global level.
Its conclusions recommend that governments establish a political dialogue within the framework of the United Nations, in which other actors involved in the development of AI can intervene, with the aim of developing regulatory frameworks based on human rights. The Committee recalls that the new standards created must be “interoperable” in different countries to avoid creating black holes in the supervision of machine learning algorithms.
They also suggest creating a “global AI capacity building network,” facilitating access to training, as well as the computing resources and datasets needed to study these systems. This network would be supported by a global AI fund to address skills gaps, prevent the formation of a major technological knowledge gap, and encourage local efforts to achieve the Sustainable Development Goals.
The report will also serve as a starting point for what is called Future Summita UN meeting that will take place between September 20 and 23. “Rapid advances in technology and science hold the promise of a better, more peaceful and prosperous future for people and the planet. However, in our rapidly changing and interconnected world, we are not keeping pace with new risks and opportunities,” the supranational organization justifies.
The aim of the summit will be to ratify a “Pact for the future chart a course towards achieving the Sustainable Development Goals and responding to new challenges and opportunities. The compact will include as annexes a Digital Global Compact and a Declaration on Future Generations,” the UN explains.
