‘A Human in the Machine’: UN AI Panel Begins Landmark Study on Global Impact of Artificial Intelligence
UNITED NATIONS, New York — The UN’s first Independent International Scientific Panel on Artificial Intelligence—a global body of 40 world-leading experts—is gearing up for its inaugural in-person summit, tasked with navigating the volatile intersection of innovation and ethics. The panel, formally appointed by the General Assembly in February, is launching a landmark study into the forces transforming modern life, from labor markets and health systems to the very nature of human decision-making.
“We are not just focusing on AI as a mathematical or algorithmic field,” said Menna El-Assady, an assistant professor at the Swiss Federal Institute of Technology (ETH) in Zurich and one of the panel’s founding members. “We are also looking at ensuring that humans are central to decision-making.”
The Role of Opinion in Journalism: Ethics, Standards & Democracy.
The panel, which brings together experts from academia, the private sector, civil society, government, international organizations, and the technical community, represents the first global scientific body dedicated to assessing how AI is reshaping our world. Its members have backgrounds in core technical AI expertise; applied AI, safety, and infrastructure experience; and AI policy, ethics, and impact. The panel is mandated to produce an annual report with evidence-based scientific assessments related to the opportunities, risks, and impacts of artificial intelligence—reports that will be presented at the United Nations Global Dialogue on AI Governance.
Augmented Intelligence: Humans and Machines Together
El-Assady, an Egyptian national, has developed the concept of “augmented intelligence” —using AI to enhance human capabilities rather than replacing humans altogether. This approach, which she has pioneered at ETH Zurich, focuses on building cooperation between AI and people across various fields, from medicine to education to public administration.
White Salt vs Pink Salt: Which Is Actually Better for Your Health? | Why Sleep Is the Most Powerful Health Habit You’re Probably Ignoring.
“We are trying to work out when we need to rely on humans and their expertise, and when things can be automated,” she explained. “We need to understand the link between AI and human models, what’s known as the co-adaptation loop, and the evolution that occurs whenever humans receive new information, or when AI does.”
The concept of “a human in the machine” is often invoked in discussions of AI ethics. It refers to the idea that a human should always be involved in decisions made by AI tools—that algorithmic outputs should be reviewed, contested, and contextualized by human judgment, especially when those decisions affect people’s lives, rights, or livelihoods.
For the UN panel, this principle is foundational. The group is not merely examining AI as a technical field but is assessing how AI systems are being deployed in real-world contexts—and whether those deployments are serving human flourishing or undermining it.
A Panel for the World, Not Just for Tech Hubs
One of the panel’s key mandates is to ensure that the global conversation about AI is not dominated by a handful of wealthy countries or tech corporations. El-Assady emphasized the need for a “public digital infrastructure” that would provide everyone who wants to develop AI with the resources they need—whether computing power, data sets, or training.
“We also need to look at how to incorporate different cultures and languages within AI models so that they are not limited to a small number of countries,” she said.
Currently, most large language models and AI systems are trained primarily on English-language data, often from Western sources. This means that the resulting tools may not work well for speakers of other languages, may not reflect non-Western cultural norms or values, and may perpetuate biases that are particular to the data sets on which they were trained.
The panel’s work aims to address these disparities, ensuring that the benefits of AI are shared more equitably and that the risks are not borne disproportionately by marginalized communities.
Trust, Ethics, and the Frankenstein Warning
The launch of the panel reflects growing concerns about the risks of unregulated AI—concerns that have been voiced at the highest levels of the UN. In September 2025, UN Secretary-General António Guterres warned the Security Council that “humanity’s fate cannot be left to an algorithm.” And in February, Volker Türk, the UN High Commissioner for Human Rights, cautioned that AI developers building models without an understanding of fundamental social and ethical principles risk creating “Frankenstein’s monster.”
El-Assady shares these concerns. She argues that ethics and trust are crucial to the sector, as is an awareness of the limitations of AI models. “AI systems are not magic,” she said. “They are tools. And like any tool, they can be used well or poorly. The question is whether we have the governance structures in place to ensure they are used well.”
One possible solution she has put forward is “AI watermarking” —a technical method to make it clear whether content has been human-originated or AI-generated, and to help users distinguish between the two. Watermarking could help combat disinformation, protect intellectual property, and maintain trust in digital content. It is one of the many topics that could be included in the panel’s first report.
From Principles to Practice: The First Report
The panel’s inaugural report is due to be released at the Global Dialogue on AI Governance, which takes place on 6–7 July in Geneva. That report will not be a set of rules or regulations—the panel is explicitly not a regulatory body, and it will not set standards or prescribe policy. Instead, it will provide rigorous, evidence-based, policy-relevant but non-prescriptive analysis, designed to inform decision-making by governments, international organizations, the private sector, and civil society.
The report will examine how AI is affecting different areas of society, including:
- The labor market: Which jobs are most at risk of automation? Which new jobs are being created? How can workers be supported through transitions?
- The health system: How can AI improve diagnosis, treatment, and public health surveillance? What are the risks of algorithmic bias or data privacy breaches?
- Education: How can AI be used to personalize learning and expand access? What are the risks of deepening digital divides or undermining human teaching?
- Democratic governance: How can AI be used to improve public services and citizen engagement? What are the risks of surveillance, manipulation, or erosion of trust in institutions?
- International peace and security: How are AI systems being integrated into military and intelligence operations? What guardrails are needed to prevent autonomous weapons or escalation spirals?
The Governance Gap
One of the central challenges the panel will address is the governance gap—the fact that AI technology is developing far faster than the laws, norms, and institutions meant to regulate it. While there are emerging frameworks, such as the European Union’s AI Act, there is no global treaty or international body with the authority to set binding standards for AI development and use.
The UN panel is not intended to fill that gap directly. But by providing authoritative, scientific assessments of AI’s impacts, it can inform and accelerate efforts to build governance at national, regional, and global levels.
“We are trying to work out when we need to rely on humans and their expertise, and when things can be automated,” El-Assady said. “That is not just a technical question. It is a social, ethical, and political question. And it requires input from everyone.”
A Diverse Panel for a Global Challenge
The panel’s 40 members reflect the diversity of expertise and perspective needed to tackle such a complex issue. They come from different regions, different sectors, and different disciplinary backgrounds. Some are computer scientists; others are economists, lawyers, philosophers, or sociologists. Some work in academia; others in government, civil society, or the private sector.
This diversity is intentional. AI is not just a technical challenge—it is a human challenge. Its impacts will be felt differently in different places and by different communities. The panel’s composition is designed to ensure that those differences are taken into account.
El-Assady, for her part, brings both technical expertise and a global perspective. Born in Egypt and educated in Germany and Switzerland, she has worked on AI ethics and human-computer interaction across multiple cultural contexts. Her concept of augmented intelligence—using AI to enhance human capabilities rather than replacing humans—offers a vision of AI that is collaborative rather than adversarial.
“We need to understand the co-adaptation loop,” she said. “Humans and AI are evolving together. The question is whether that evolution will be intentional or accidental, equitable or exploitative, democratic or dictated by a few powerful actors.”
What Comes Next?
The panel’s in-person summit will be the first opportunity for members to meet face-to-face and begin the substantive work of drafting the inaugural report. Between now and July, the panel will review existing research, commission new studies, consult with stakeholders, and deliberate on findings.
The July report will not be the last word. The panel is mandated to produce annual reports, building a growing body of evidence over time. As AI technology continues to evolve, the panel’s assessments will need to evolve with it.
“This is just the beginning,” El-Assady said. “But it is an important beginning. For the first time, we have a global scientific body dedicated to understanding AI’s impacts and ensuring that humans remain at the center. That is a big step forward.”
SOURCES / INPUTS
https://news.un.org/en/story/2026/04/1167263
The Future of Science & Technology: AI, Space, Biotechnology & Digital Transformation Explained.
France Moves Closer to Social Media Ban for Under-15s as Senate Approves Conditional Restrictions | Indonesia Begins Implementing Landmark Social Media Ban for Children Under 16.
