The top technology official of the United Nations talks about artificial intelligence, fostering global unity, and the concerns that cause him sleepless nights.

UNITED NATIONS — Artificial intelligence, and how and whether to regulate it, has gotten a lot of discussion in and around this year’s U.N. General Assembly meeting of world leaders. With a U.N. advisory group on AI set to convene this fall, the world organization’s top tech-policy official, Amandeep Gill, sat down with The Associated Press to talk about the hopes, concerns and questions surrounding AI.

Here are excerpts from the interview, edited for length and clarity.

___

What unique contributions can the United Nations bring to the discussion and regulation of artificial intelligence, considering that several national governments and multinational organizations are already engaged in this matter?

GILL: I would choose three terms. Inclusiveness – which means bringing together a larger number of countries, in comparison to other significant initiatives. The second term is legitimacy, as the U.N. has a history of assisting countries and other entities in handling the effects of various technologies such as biology, chemistry, nuclear, and space science. This assistance includes preventing misuse and promoting the inclusive and peaceful utilization of these technologies for the benefit of all.

The third aspect is authority. When the U.N. takes a stance on something, it carries weight and influence. The U.N. has specific mechanisms, such as human rights treaties, that can be connected to these commitments. For instance, if an AI feature results in the marginalization of a particular community or the infringement of certain people’s rights, governments are obligated, based on the treaties they have ratified at the U.N., to prevent such occurrences. Therefore, it is not solely a matter of moral authority; it also generates pressure to adhere to the commitments made.

Are there challenges that the U.N. faces that other active entities do not face, or not to the same extent?

GILL: To successfully manage a large group, it is important to have an effective process that goes beyond simply having everyone present. It should involve meaningful discussions and lead to positive results. Another challenge is to actively involve the private sector, civil society, and the technology community. That’s why the Secretary General’s advisory body on AI governance is being formed as a multi-stakeholder body, with careful consideration.

One more restriction is that the processes of the United Nations can be time-consuming due to the need for consensus among numerous participants, while technology advances rapidly. Hence, it is necessary to enhance our agility.

Is it possible for governments, regardless of their level, to effectively regulate AI?

GILL: Absolutely. I believe governments have the ability to shape the path of AI in various ways. It involves more than just implementing regulations to prevent misuse and harm, but also safeguarding democracy and the rule of law. Additionally, it entails fostering an innovation ecosystem that is diverse and inclusive, reducing the concentration of economic power and ensuring equal opportunities for all.

AP: When it comes to equal opportunities, individuals in the Global South are optimistic about the potential of AI to bridge the digital gaps. However, there is also apprehension that specific nations may gain advantages from this technology while others are excluded and left behind. Do you believe it is feasible for everyone to align their perspectives and experiences?

GILL: I completely agree that this is a crucial concern. It is an opportunity for us all to unite in a more sophisticated manner, moving beyond the simplistic view of “promise and peril” that is often held by those who have the ability to take action. Instead, we should focus on a more comprehensive understanding that includes access to opportunities and the empowerment it brings, in addition to the potential risks involved.

Yes, indeed, there exists an opportunity and a sense of excitement. However, determining how to take advantage of this opportunity is an incredibly significant inquiry.

AP: There is significant discussion surrounding the idea of consolidating global conversations regarding the regulation of AI. What is your interpretation of this concept, and how do you believe it can be achieved?

A: It would be highly significant to reach a convergence and mutual understanding regarding the risks. It would also be extremely valuable to have a shared understanding of effective governance tools, as well as areas that require further research and development. Similarly, a unified understanding of the necessary agile and decentralized model for governing AI, with the aim of minimizing risks and maximizing opportunities, would be of great importance. Lastly, achieving a common understanding of the political decisions to be made at the Summit of the Future in September 2024 is crucial for ensuring the sustainability of our efforts and gaining the trust and support of the public.

AP: What concerns you the most about AI during nighttime? And what brings you optimism when you start a new day?

GILL: I want to begin with the positive aspect. I am truly enthusiastic about the opportunity to enhance advancements in the Sustainable Development Goals through the use of AI, specifically in crucial fields like health, agriculture, food security, education, and the green transition. However, my concern lies in the possibility of allowing AI to progress in a manner that, firstly, misleads us regarding its capabilities, and secondly, results in a greater consolidation of technological and economic influence in the hands of a select few. Although these individuals and companies may have good intentions, democracy flourishes through diversity, competition, and openness.

I am optimistic about AI in general, but I do have concerns about it being used to undermine democracy, deceive society, and diminish our humanity. These are the issues that worry me, and I hope we navigate AI in the right direction.