WITH little fanfare, the United Nations has assembled a team of 40 leading experts from different backgrounds to assess how artificial intelligence is transforming modern life and establish the framework to ensure that the transformation aligns with ethical and societal values.
The UN’s Independent International Scientific Panel on AI is the first global body of its kind. Its creation “reflects the growing concerns about the risks of unregulated AI,” the UN said.
In September last year, UN Secretary General Antonio Guterres warned that “humanity’s fate cannot be left to an algorithm.” Volker Turk, the UN High Commissioner for Human Rights, voiced the same fear.
“It reminds me a little bit of Frankenstein’s monster; you develop something that you don’t control anymore,” Turk said in one interview.
AI undeniably offers a plethora of benefits. It can perform tasks with the speed and accuracy humans could never match. It can analyze vast amounts of data and identify patterns and trends faster, vastly improving decision-making critical in fields like medicine, health care and financing. In high-risk industries like mining and construction, it can perform tasks too dangerous for human workers.
But there are downsides as well. AI can lead to job losses, as automation renders some tasks obsolete. An overreliance on AI increases the risk of systems failure. There have been documented instances of what is known as “hallucinations,” when AI systems “produce outputs that are plausible sounding but factually incorrect or nonsensical,” according to the World Bank report.
“These errors occur because the AI generates text based on patterns and data it has been trained, without an understanding of the real-world context or factual accuracy,” it said.
AI could spawn ethical issues like data privacy, surveillance and biased decision-making. As AI becomes increasingly integral to modern society, the need to effectively govern it grows more urgent.
“Ensuring that AI is developed and deployed in a manner that is ethical, transparent, and accountable requires robust governance frameworks that can keep pace with technological evolution,” the World Bank noted in its 2024 report, “Global Trends in AI Governance.”
The multilateral agency believes that AI governance “cannot rely on a single, universal approach, and no regulatory model works in isolation.” It underlined the importance of “adopting a flexible, adaptable governance framework that evolves with both technological advancements and societal changes.”
Each country “must assess the maturity of its AI ecosystem, existing legal and regulatory landscapes, and available resources” in determining which approach to embrace.
A one-size-fits-all approach “is unlikely to work given the diversity of AI applications and risks,” the World Bank said.
It cautioned policymakers against “importing” regulatory provisions or strategies from other countries “without appropriate modifications and consultation with affected communities, the public, civil society, the private sector, and the international community.”
One particularly provocative suggestion by the World Bank is setting up a “regulatory sandbox,” a “controlled and time-based environment” for developing and testing new products and technologies before they are released into the market.
Regulatory sandboxes “enable government agencies to maintain oversight and control while creating a dynamic regulatory environment for testing emerging technologies and business models.”
In navigating the complex environment of AI governance, the UN panel will be guided by the concept of “augmented intelligence.” One of the experts, Menna El-Assady, describes it as using AI to enhance human capabilities rather than replacing humans altogether.
Augmented intelligence is built around the slogan, “a human in the machine.”
El-Assady is calling for a “public digital infrastructure” that anyone who wants to develop AI can tap into. “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.
The UN made it clear that the panel is not a regulatory body, “and will not set rules, enforce standards, or prescribe policy.” What it will do is “provide the rigorous, evidence-based, policy-relevant but nonprescriptive analysis that informed decision-making requires.”
Turk has high hopes that, properly regulated, AI can be a “fantastic tool” that could usher in “an inclusive, meaningful, participatory type of development, that helps us solve the many problems and challenges in today’s world.”