Who will control the internet?

Earlier this month, a group of more than 30 prominent digital pioneers and engineers fired off a distress signal, warning that negotiations underway at the United Nations threatened something of a hostile takeover of the internet by governments seeking greater control over a technology that has transformed the world in a single generation. The group, which includes Vint Cerf, one of the fathers of the internet, and Tim Berners-Lee, founder of the World Wide Web, expressed concern about the pace of negotiations over a draft U.N. Global Compact, which aims to write the rules of the road for digital technology. The compact — being negotiated in U.N. talks led by Zambia and Sweden — is expected to be endorsed by world leaders at the Summit of the Future in late September. “Some proposals for the Global Digital Compact (GDC) can be read to mandate more centralized governance,” the group wrote in an open letter to U.N. Secretary-General António Guterres and his tech envoy, Amandeep Singh Gill. “We are concerned that the document will be largely a creation of governments, disconnected from the Internet and the Web as people all over the world currently experience them.” The letter-writing campaign was triggered in part by a set of amendments to the compact advanced last month by China and a bloc of more than 130 countries from the global south, known as the Group of 77, outlining a vision in which governments at the U.N. will play a “leading role” in decisions about how the internet is governed. Those governments — which include countries like Cuba, Iran, Saudi Arabia, and Venezuela — also advocate downgrading the role of critical forums, such as the World Summit on the Information Society, or WSIS, and the Internet Governance Forum, which have involved members of the digital technical community in setting the terms for a more decentralized digital cooperation, according to a copy of amendments reviewed by Devex. The concern is that top-down, state-centric governance of the internet will stifle innovation and economic opportunities, and weaken human rights standards. There is also fear that the global compact fails to adequately affirm the need for states to refrain from the use of mass surveillance, promote digital privacy, including end-to-end encryption, and uphold international humanitarian law. The digital expert group protested that the U.N. negotiations have been largely dominated by states, providing a secondary role for the digital pioneers who have managed the evolution of the networked world from the grassroots. In contrast, the U.N. negotiations are “being developed in a multilateral process between states, with very limited application of the open inclusive and consensus-driven methods by which the Internet and Web have been developed to date. Beyond some high-level consultations, non governmental stakeholders (including internet technical standards bodies and the broader technical community) have had only weak ways to participate in the GDC process.” For the time being, governments remain deadlocked over the compact, with several governments, including China, Russia, and the United States, rejecting the draft, and setting the stage for a new round of negotiations over the coming weeks and months. Western diplomats assure that it is unlikely that China and the G77 will succeed in muscling through a final pact that fully mirrors their vision of a state-controlled internet governance regime. But as key players in the U.N. negotiations, they will exercise considerable influence over the final outcome. The U.N., meanwhile, has insisted that it envisions an inclusive internet governance process that grants a broad range of players — including internet experts, human rights advocates, businesses, and governments — a seat at the decision-making table. “Equity demands that more voices play meaningful roles in decisions about how to govern technology that affects us all, including beyond the current concentrations of decision-making in the AI technology sector,” according to a draft report, seen by Devex, by a U.N. advisory board, known as the high level body on AI, or HLAB. The legacy thing The U.N. bid to expand its role in digital governance dates back to 2020, when world leaders marked the organization's 75th anniversary by calling on Guterres to detail a vision for an updated multilateralism capable of addressing “current and future challenges.” The following year, Guterres released his reform report, titled “Our Common Agenda,” which outlines a road map for the U.N. over the next 25 years, including a proposal to adopt a Global Digital Compact. Guterres established the post of the U.N. tech envoy in January 2021, part of an effort to map out a role for world organizations in harnessing digital technology to end world poverty and global inequality, and to address the risks posed by AI. Gill, a former Indian diplomat who serves as the current envoy, was appointed in June 2022. In October 2023, Guterres established HLAB, to begin “a global, multidisciplinary, multistakeholder conversation on the governance of AI so that its benefits to humanity — all of humanity — are maximized, and the risks contained and diminished.” The panel — which is comprised of 39 members, including the U.N. tech envoy — recently circulated a draft summary of their final report, titled “Governing AI for Humanity.” They hope it will be incorporated into the Global Digital Compact, which is reaching the final stages of negotiations. “Amandeep released the report [to key players] in an effort to influence the GDC negotiations,” said one observer. Those negotiations, the source said, “are in the final sprint” and the panel “wants to make sure they are maximizing the impact, before it’s too late.” “Equity demands that more voices play meaningful roles in decisions about how to govern technology that affects us all.” --— From the draft report by the U.N. High-Level Advisory Body on Artificial Intelligence “Frankly, it is all really crazy and things are moving too fast,” the source added. But the U.N. has struggled to allay suspicions that governments, including those like China, Russia, and Saudi Arabia, will dominate discussions at the U.N. Experts cite a trust deficit that feeds concerns that the U.N. is seeking a power grab. “The U.N. tech envoy wants to seize control of the process and centralize governance in the U.N.,” said Konstantinos Komaitis, a nonresident fellow at the Atlantic Council, and a signatory to the letter to the U.N. leadership. “The point is that in complex systems like the internet and AI, centralization is a recipe for disaster. You need the participation of different people. You need to create a structure that effectively creates accountability for the way decisions are made.” AI cries out for governance The internet spat comes as HLAB has circulated an ambitious plan that would place Guterres at the center of the governance of artificial intelligence. It proposed the creation of a “light, agile” AI office housed in the U.N. secretary-general’s office in New York that would manage a scientific panel based loosely on the Intergovernmental Panel Climate Change, or IPCC, to serve as a clearing house for leading-edge research that could help educate scientists, policymakers, and governments. It would maintain a registry of internet standards, establish an AI training data initiative, and set up a global fund to help underwrite the costs of transferring expertise and technology to countries in the global south. The panel initially proposed a start-up fund of $1 billion, but reduced it to $100 million. “These systems cry out for governance,” the advisory panel’s report stated. “There is, today, a global governance deficit with respect to AI.” But U.N. member states balked at the drive to grant a central role to the U.N. secretary-general in AI governance — though they did agree to establish a scientific panel and a global policy dialogue. The latest draft of the Global Digital Compact, which was reviewed by Devex, scrapped the proposal for the AI fund, instead asking the U.N. chief to propose new financing options. It also kicks the proposal for a new AI office down the road, requesting Guterres, in consultation with governments, propose a detailed plan next year for a U.N. office to coordinate the organization’s AI activities. But it alarmed civil society advocates. On Friday, a coalition of digital rights organizations published a statement sharply criticizing the latest draft of the global compact, saying it centralized too much power in the hands of the U.N. secretary-general and the language was too weak on privacy, human rights, and gender equality. The draft compact’s provisions on human rights are “not sufficiently robust” and it “fails to recognize the need for States to refrain from the use of mass surveillance and ensure that targeted surveillance technologies are only used in compliance with international humanitarian law.” The government trust deficit David Kaye, a law professor at the University of California, Irvine who served as the U.N. special rapporteur on the promotion and protection of the right to freedom of opinion and expression, said there is nothing inherently wrong with the U.N. seeking to carve out a role as a clearing house for the latest scientific developments on AI. “That’s not a terrible idea,” he told Devex in a telephone interview. But he voiced concern that centralizing AI governance efforts at U.N. headquarters, a highly political venue, will inevitably inject politics into technical decisions over the way digital networks communicate across borders. “I don’t think the tech envoy and others in the U.N. want to undo that technical interoperability, but the more you put this into the U.N., the more you naturally move into a state-dominated environment,” Kaye added. “The internet only works across borders when the technology allows that to happen, right? But countries like China and Russia have technology to block the ability of the internet to work across borders.” Kaye said the technical dissenters “are concerned that the traditional mode of internet governance at the technical level, which is non-state driven, is kind of slowly being pulled into this process being driven by states and states making these decisions. I think they have a totally legitimate concern that if the norms move toward a state-centric approach the traditional way of thinking about ensuring interoperability across networks, across the internet, will slowly deteriorate, and that's a problem for the global internet.” Mission creep? The U.N. chief’s bid to carve out a central role in internet and AI governance has proven controversial with key governments, including the U.S., which has expressed concern that the U.N. is creating an unnecessary new bureaucracy to fulfill tasks that are already being undertaken by a number of U.N. institutions. It has also opened turf battles with U.N. agencies, including the International Telecommunications Union and UNESCO. Their supporters suspect Gill is seeking to exercise greater authority over those agencies. “The concern is mission creep,” said one observer. “You hear multiple people saying the secretary-general is looking to shore up his legacy and Amandeep is looking for a job.” The proposals will feed into negotiations among states on a digital compact. There is considerable pushback. The scientific panel, and the policy dialogue, are being “heavily litigated” in the latest draft, according to one diplomatic source. There is no agreement on a fund, let alone on a figure. Even the proposal for a new AI office in the secretariat is being kicked down the road. The latest draft merely requires Guterres to submit a report to the U.N.’s budget committee, outlining the costs and staffing — “so this is all TBD,” the source said. The ambitious new proposal contrasts with the U.S. calls for a more measured role for the U.N. in creating new institutions to manage artificial intelligence, noting that several U.N. agencies already play a role in setting technical and ethical standards for AI. “It would be premature to call for establishing new U.N. governance mechanisms without a clearer understanding and fulsome consensus on where there may be gaps in the ability of existing U.N. agencies to address AI,” according to a confidential U.S. paper shared with member states. “It is our view that any proposals for new processes, panels, funds, partnership, and/or mechanisms are premature.” The U.S., meanwhile, has been vying with China to claim the mantle of leadership in shaping the U.N.’s role in harnessing artificial intelligence to address global development challenges. On July 2, the U.N.’s 193-member General Assembly adopted a Chinese resolution urging governments to “bridge the artificial intelligence and digital divides” between wealthy and lower-income countries, and to harness the new technology to accelerate progress on the U.N. Sustainable Development Goals, which set a series of ambitious targets, including the elimination of poverty, by the year 2030 — targets which are all but certain not to be met. The Chinese resolution — which was sponsored by more than 140 countries — was passed by consensus. The U.S. initially had misgivings about the Chinese initiative, but backed it after Beijing offered a series of concessions: It underscored the importance of promoting human rights, stripped out provisions denouncing the use of unilateral sanctions, and amended a provision calling on the U.N. to play a central role in AI governance. The U.S. has also sought to shape the debate on artificial intelligence at the U.N., leading negotiations on the first resolution adopted in March by the General Assembly. The resolution, sponsored by more than 120 countries, and supported by China, vows to build “global consensus on safe, secure and trustworthy artificial intelligence systems” that advance progress on the U.N.’s SDGs. That text “will serve as a foundation for multilateral AI efforts and existing and future U.N. initiatives.” In the end, the U.N. member states — which never gave Guterres an explicit mandate to erect an AI governance system, let alone pay for it — may simply scale back the U.N.’s ambitions. Some governments may ultimately request the U.N. flesh out its plans and “see if they can kick the can down the road,” according to one U.N.-based observer.