The year 2026 marks the beginning of the second quarter of the first century in the third Millenium.

When the 21st century started, there was a lot of hope. The Internet removed the old barriers of time and space for the distribution of information. The new freedom of communication fed the hope for more democracy, sustainable economic growth and new job opportunities. The dream of a “new home of mind” was the driving force of the early days of the information revolution. “We are creating a world that all may enter without privilege or prejudice accorded by race, economic power, military force, or station of birth. We are creating a world where anyone, anywhere may express his or her beliefs, no matter how singular, without fear of being coerced into silence or conformity,” wrote John Perry Barlow in his “Declaration of Cyberindependence” in Davos in 1998. In 2000, there was no Facebook, no Instagram, no TikTok and no smartphone. “Don’t be evil” was the motto of Google, a young search engine, established in 1996 by two PhD students at Stanford.

25 years later, there is a lot of fear. The promises of yesterday have turned many hopes into their contrary. The risk is now high that the world is moving towards a war with Internet-based lethal autonomous weapon systems (LAWS). The free Internet is misused to undermine democracy. The benefits of the digital economy are unequally distributed and have deepened the global divide between the poor and the rich. Human rights, such as freedom of expression and privacy, are under pressure from growing censorship and mass surveillance. Bad guys, such as cybercriminals, hate preachers, and liars, have occupied large parts of cyberspace. Millions of people, and not only youngsters, are slaves of their technical devices and “amusing themselves to death”, as Neil Postman predicted in 1985. The information revolution is eating its children.

However, the successful outcome of the WSIS+20 review has lit some candles of hope in the global digital environment. Will 2026 become a good year? Will the global Internet community be able to find the right path between fear and hope into a “people-centered, inclusive and development-oriented information society”, as it was reiterated as a vision for the future by the WSIS+20 Outcome Document in December 2025?

Let’s have a look at the five baskets of the global Internet Governance Ecosystem: Cybersecurity, digital economy, human rights, artificial intelligence and the management of critical Internet Resources.

1. Cybersecurity

Cybersecurity isn’t primarily a question of the security of networks and devices anymore. It is a question of national security. And it can become an issue of war or peace. The wars of the last years—Ukraine, Gaza, Sudan, Iran and others—have triggered the development of a new generation of Internet-based weapon systems. We are in the midst of a digital arms race. The Internet is now fully integrated in all kinds of military operations, both for defence and offence. Cyberattacks against infrastructure and disinformation campaigns over social networks are part of the new hybrid warfare. And cybercrime is growing. In its annual Global Risks Report, the World Economic Forum (WEF) has put cybersecurity again among the main five threats.

1.1 UN Negotiations

Cybersecurity has been discussed in the 1st Committee of the UN General Assembly (UNGA) since 1999. In 2015, a group of governmental experts (GGE) proposed 11 voluntary norms for responsible state behaviour in cyberspace, which were later confirmed by the UN. The eleven norms are seen today as the most relevant cybersecurity framework in international law. In 2020, an open-ended working group (OEWG) began discussing cybersecurity issues more systematically, including cybernorms and principles, the relevance of international law in cyberspace, and confidence—and capacity-building measures. In 2025, the OEWG was transformed into a permanent “Global Mechanism”.

But all the diplomatic exchanges did not stop the digital arms race. Billions of dollars are invested in a new generation of Internet-based weapon systems. Drones are the tanks of the new generation of wars. And this risk of an AI war with killer robots is growing.

Nevertheless, the OEWG did produce something. It agreed on a so-called “Point of Contact Directory” (PoC). The PoC is like a red telephone, where governments have a number to call if they see a suspicious cyberattack. This is a small but good confidence-building measure. Furthermore, the OEWG launched an Online Portal for Cybersecurity Capacity Building. This will help in particular countries from the Global South to enhance their national capacities to deal with cyberattacks.

The first meeting of the new “Global Mechanism” will take place in New York in March 2026. But it would be unrealistic to expect that the new mechanism will pave the way towards a breakthrough in digital disarmament. If we see some more confidence and capacity-building measures, this would already be a success.

A similar “limited progress” can be seen with regard to lethal autonomous weapon systems (LAWS). For more than ten years, the issue has been discussed, both under the umbrella of the Convention on Conventional Weapons (CCW) and in the 1st UNGA Committee. It is widely recognized that this is a burning issue. Killer robots are seen by a majority in the world as a nightmare. The late Pope Francis called for a ban on LAWS. UN Secretary General Guterres called for a legally binding instrument to regulate LAWS. A group of academics, including members of the Internet Hall of Fame, did send in August 2025 a “Potsdam Call on Autonomous Weapon Systems” to the president of the 80th UNGA, calling for an end of the AI arms race by refering to the 80th anniversary of the Potsdam Conference, which ended World War Two, and the Einstein-Russel Manifesto from 1955, which called for an end of the nucelar arms race.

But for years, the LAWS discussion has been moving in circles. The only outcome so far is an agreement on a set of guiding principles and on a “two-tier approach”: LAWS, which are out of human control, should be prohibited as chemical or biological weapons. And the use of LAWS under human control should be linked to the rules of the Geneva Conventions on International Humanitarian Law. In December 2025, the 80th UNGA adopted a resolution (80/57) which recognized “substantial progress” on a “rolling text” and invited member states to “complete the set of elements for an instrument with a view to future negotiations.” But even this vague language was not adopted by consensus. 164 states voted in favor, six voted no and seven abstained. Under the “No” votes were the USA, Israel, Russia and North Korea.

Another UN Resolution (80/58) dealt with the role of AI in the military domain in a more general way. The resolution states that there are both opportunities and risks in military AI applications and proposes “to deepen the understanding of the humanitarian, legal, security, technological and ethical perspectives.” It was decided to “convene informal exchanges” in Geneva to further “share perspectives on observations and conclusions.” The three-day meeting will take place in Geneva in the summer of 2026. The resolution got 167 Yes-votes, five No-votes, including USA, Israel, Russia and North Korea, and five abstentions.

The UN Resolution 80/23 on “Possible Risks of the Integration of AI into Command, Control and Communication Systems of Nuclear Weapons” got nine No’s. The club of the four Non-Nuclear-Weapon States was joined by other nuclear powers, such as France, UK and China. Also, Argentina voted No. The resolution warns of giving decision-making capacity to AI in managing nuclear arsenals. This could raise the “risk of accidental, unintended or unauthorized use of nuclear weapons”. In a worst-case scenario, the world could become the victim of an AI hallucination. Is this a wake-up call? Anyhow, the hope is that as long as governments talk to each other, they will avoid using force against each other.

A new element in this growing complexity of cybersecurity discussions is the role of non-governmental actors in the military domain. There is a growing need for closer cooperation among the military, the private sector and the technical community. The “Global Commission on Responsible Use of AI in the Military Domain” (GC REAIM) recommended in its Final Report from July 2025 that a multistakeholder approach is needed in the military field. Resolution 80/58 got inspired by this proposal. The “informal exchange” in Geneva in summer 2026 will be open to all stakeholders. Is this another reason for hope? But there is also a fear that new AI-based weapon systems will end up in the hands of private actors with bad intentions, such as terrorist groups or organized crime.

1.2 The Hanoi Convention

An important part of cybersecurity is the fight against cybercrime. Only the financial losses go into the billions of Dollars annually. In October 2025, the new UN Convention against Cybercrime was opened for signature. The signing ceremony for the new UN Convention, now called the “Hanoi Convention”, took place in the capital of Vietnam on October 25, 2025, in the presence of UN Secretary-General Antonio Guterres. Vietnamese Prime Minister Pham Minh Chinh welcomed more than 100 governments and numerous non-governmental representatives. Guterres said at the opening: “In cyberspace, nobody is safe until everybody is safe. One vulnerability anywhere can expose people and institutions everywhere. That is why we need a strong, collective, global response… The UN Cybercrime Convention is a powerful, legally binding instrument to strengthen our collective defences against cybercrime. It is a promise that fundamental human rights such as privacy, dignity, and safety must be protected both offline and online. It is a testament to the continued power of multilateralism to deliver solutions. And it is a vow that no country, no matter its level of development, will be left defenceless against cybercrime.”

By the end of 2025, 72 governments out of 193 UN states had signed the convention. Russia initiated the negotiations in 2019 and, like China and numerous developing countries, it was one of the first signatories. The European Union also signed the convention, but more than half of the EU member states were hesitant to sign at this stage. The US, which had voted in favour of the convention at the 79th UNGA in December 2024 (before the Trump administration took office), attended the signing ceremony but declined to sign the convention. Instead, the US spokesperson invited governments to join the ‘Global Coalition on Fostering the Future Together’ presented by First Lady Melanie Trump at the opening of the 80th UNGA in September 2025. The Hanoi Convention will enter into force 90 days after the 40th instrument of ratification has been deposited.

Western reservations about the UN convention are related to the vague definition of cybercrimes, the lack of safeguards against violations of human rights such as freedom of expression and privacy, and international extradition procedures that are open to abuse. The offer by Western states to join the Council of Europe’s 2001 Budapest Convention on Cybercrime was rejected by the majority of countries in the Global South because they were not involved in the drafting of this convention. However, negotiations by the relevant Ad Hoc Committee (AHC) succeeded in establishing fundamental consistency between the Budapest Convention and the new Hanoi Convention.

In addition to over 100 governments, the Council of Europe, Interpol and non-governmental stakeholders were also represented at the signing ceremony. Google’s Legal Council Nima Binara spoke on behalf of many Western observers: “For Google, the protection of human rights online is inseparable from our mission. Google believes the following three areas should feature strongly in any discussion: First, the continued, robust involvement of multistakeholders is vital at all further stages of the Convention. Second, it’s vital that substantive human rights protections in the Convention be strengthened. Third, we believe it’s equally vital that procedural safeguards should be strengthened.”

1.3 National Efforts

But the hope that international agreements will produce more cybersecurity is limited. More and more countries are taking practical and regulatory measures to strengthen their cybersecurity on the national level. “My country first” is also the slogan for security in cyberspace. And cybersecurity is increasingly taking on a military component.

The US is working on a new National Cybersecurity Strategy, which will be published in early 2026. According to leaked information, the strategy will contain six chapters: 1. cyber offence and deterrence; 2. aligning regulations to make them more uniform; 3. bolstering the cyber workforce; 4. federal procurement; 5. critical infrastructure protection; and 6. emerging technologies. The new strategy is primarily aimed at warding off cyber attacks from China. The White House National Cyber Director, Sean Cairncross, stated at the Meridian Summit in Washington in October 2025: “To date, I don’t think the US has done a terrific job of sending the signal, in particular to China, that their behaviour in this space is unacceptable. It is meant to harm us. It sits on our critical infrastructure systems and threatens chaos. It tries to put us in strategic dilemmas that impact our decision-making. And that is something that is scaling. It is something that is seen as cost-free, I think, across the ocean, and that is something that needs to be reset so that there is strategic stability in this domain.”

China amended its Cybersecurity Law in the fourth quarter of 2025. It came into force on January 1, 2026. A key addition to the new law is the stricter reporting requirements for companies. Operators of critical information infrastructures are now required in certain cases to report significant cybersecurity incidents within just 60 minutes. In other cases, the reporting deadline is extended to four hours. These requirements are reinforced by administrative measures for reporting national cybersecurity incidents, which were published by the Cyberspace Administration of China (CAC) on November 1 2025. This means that there is now a uniform framework that applies to all network operators who set up networks within China or offer services via Chinese networks. Penalties for non-compliance with reporting requirements can be as high as €1 million.

On October 16, 2025, the European Commission adopted a Defence Readiness Roadmap that supplements existing cybersecurity regulations such as NIS 2. European Commission Vice-President Henna Virkkunen said: “Disruptive technologies and their rapid development, testing and embedding in defence capabilities are essential to modern warfare. The war in Ukraine clearly demonstrates how fast defence technologies evolve and how frontier technologies such as AI, quantum, cyber and space-based systems provide rapid tactical change on the battlefield.’ The roadmap aims to bring the tech and defence industries closer together and increase Europe’s production capacity through innovation.”

And also, Russia concentrates more on national efforts. At a meeting with the leadership of the Russian Ministry of Defence on December 17 2025, Russian President Putin called for rapid rearmament in the digital sphere: “Key areas of the state programme include air and missile defence systems, command and control systems, electronic warfare capabilities, and unmanned aerial vehicles (UAVs) operating in all environments… The Russian army must continue to remain at the forefront of technological progress. This means accelerating the introduction of robotics, information technologies, and new materials into the Armed Forces, as well as expanding the use of artificial intelligence in command and control systems and autonomous combat platforms.”

2. Digital Economy

One thing is for sure: In 2026, the digital economy will continue to grow. Political turbulences like problems with supply chains, unilateral sanctions and national protectionism will diminish the digital economy, but will not stop the expanded use of AI with all its new services and applications. Will the AI bubble burst in 2026? Probably. But this will not stop new AI services. Do you remember the burst of the “Dot Com Bubble” in 1999? It did not stop the digital transformation. It just cleaned an overheated market.

2.1 The Battle among Digital Empires

However, the digital economy will become more than ever a political battlefield. When Ann Bradford published her book “Digital Empires” in 2022, she recognized the US and China as the main digital players and the EU as a digital regulatory power. But this view is now fundamentally challenged. The “Brussels Effect” will not work anymore as in the past.

While China is still using European approaches to regulate the digital space (but ignores human rights safeguards), the US under Trump has fundamentally challenged the EU’s role as a digital rule maker. Donald Trump made clear that he will fire back if the EU continues to use its regulatory stick to punish US tech companies for non-compliance with its digital rulebook.

The EU is now maneuvering. Just recently, the EU sent a signal to the Trump Administration by its “Digital Omnibus”, which reduced some regulatory burdens related to the data protection law GDPR and the EU AI Act. But the main contested EU regulations, the Digital Market Act (DMA) and the Digital Service Act (DSA) were not included in the Digital Omnibus. This could lead to an open battle.

China will be watching what is going on in this EU-US battle. But there are other emerging digital empires—India, Brazil, UAE, Saudi Arabia, Indonesia, the African Union and others—which will try to get their seat on the digital table. It gets very complex.

Whether the “G20 Digital Economy Working Group” (G20 DEWG), where all the old and new digital empires are members, can play a role in finding reasonable arrangements remains to be seen. In 2026, the G20 is chaired by the US. The US has not yet published its work program for the G20 DEWG, and the dates for the meeting of the G20 digital ministers are not yet fixed. The G20 summit, however, is scheduled for December 2026 in Florida, at Trump National Doral Miami Resort.

2.2 Digital Trade, Digital Taxation and Digital Development

But there are also other efforts to find global solutions for the development of the digital economy. For years, the World Trade Organisation (WTO) has been trying to reach an agreement on a digital trade pact.

In 1999, the WTO adopted a “Moratorium on Customs Duties on Electronic Transmissions”. This agreement ensures digital trade can flow freely across borders without tariffs and added cost. It was renewed several times and has helped businesses, entrepreneurs and independent creators benefit from a borderless digital economy to compete, grow and reach customers across borders. The forthcoming 14th WTO Ministerial Meeting, scheduled for Yaounde in Cameroon in March 2026, will have to consider making the moratorium permanent. But there is also a proposal for a WTO Agreement on Electronic Commerce. After more than five years of negotiations, a text is ready. On December 3, 2025, a big group of WTO member states proposed to adopt the draft and to add it to the “Marrakesh Agreement”, the constitution of the WTO from 1994.

Digital taxation is another challenging issue. For years, the problem was discussed by a joint G20/OECD working group, which finalized its work in 2021 by the adoption of the so-called “OECD/G20 Inclusive Framework on Base Erosion and Profit Shifting” (BEPS). It agreed on a so-called two-pillar solution to reform the international tax framework in response to the challenges of digitalisation of the economy. It included a 15% minimum tax.

Implementation started slowly. However, 147 states have joined the agreement. When the Trump administration came into office in January 2025, it rejected parts of the BEPS and called for a “side-by-side agreement” with special exemption clauses for US tech giants. The G7 finance ministers accepted the US request in July 2025, and in early January 2026, the OECD announced that the member states of the Inclusive Framework had accepted the new “side-by-side mechanism”.

What does this mean? Who will be next to ask for “exemptions”?

In the meantime, another group in the UN has started work on a new “UN Framework Convention on International Tax Cooperation” in spring 2025. It established an Intergovernmental Negotiating Committee (INC) and had three meetings in New York and Nairobi. Three more meetings will follow in 2026. The plan is to finalize the UN Tax Convention by 2027.

In all those negotiations, the voice of the Global South gets louder and louder. There was also a clear signal from the ITU World Telecommunication Development Conference (WTDC) in Baku (November 2025) that it needs new efforts not only “to bridge” the digital divide, but “to close” the digital divide. “We reaffirm the importance of leaving no one behind in digital transformation, which will require urgent efforts to expand equal opportunities for all,” says the “Baku Declaration”.

Also in the WSIS+20 negotiations in December 2025, the Global South was pushing for more efforts towards digital development. It will be interesting to see what the UNCSTD Working Group on Data Governance (WGDG) will produce in 2026. Its 5th meeting is scheduled for February 2026 in Geneva. The WGDG was established under the Global Digital Compact (GDC), which has to be reviewed in 2027.

3. Human Rights

2025 was not a good year for human rights in the digital age. And there is only a little hope that 2026 will be better.

3.1 Freedom of Expression

The right to freedom of expression is misused more and more for the online distribution of disinformation and fake news, which pollutes the intellectual sphere. Internet shutdowns and censorship are now common practice by more and more governments if they see a problem with online communication. Freedom House’s “Freedom on the Net 2025” report confirmed the 15th consecutive year of global internet freedom decline, driven by authoritarian crackdowns on online dissent, particularly during protests, leading to more control and manipulation.

In her recent report to the 80th UNGA, Irene Khan, the Special Rapporteur on freedom of opinion and expression of the Human Rights Council (HRC), has also expressed her concerns with regard to the role of the big tech companies: “Against that background, a paradigm shift has occurred in the private governance of freedom of opinion and expression online. Fuelled by transformative advances in AI, the global agenda for freedom on the Internet is being reset to advance geopolitical and corporate interests in ways that could seriously jeopardize human rights, in particular freedom of opinion and expression. At stake is the equal enjoyment of fundamental human rights and the safety of the people who are the most vulnerable, the most marginalized and most subject to discrimination in society, as well as the health and survival of democracies and public trust in the integrity of information”.

The report recommends to states, inter alia: “Social media regulation should ensure that companies carry out human rights due diligence and impact assessment, align their curation and moderation policies with human rights standards and adhere to high standards of transparency and accountability.”

Irene Khan’s mandate ended on December 31, 2025. The new Rapporteur has a long to-do list.

3.2 Privacy

With regard to privacy, the situation is not much better. Law enforcement, as well as national security and intelligence agencies, are calling for more and more rights to get every data they claim to need to fight cybercrime and to deal with threats against national security. Internet users are becoming more “glass persons”, “transparent human beings”, regardless of all the data protection laws that have been adopted in recent years. Not only do autocratic regimes look into the private communication of their citizens. The US administration has recently announced that it will look into the social media accounts of US visitors when they enter US territory. Unfriendly comments could be used as an argument to deny access to the US.

New developments in AI technology will deepen and broaden those possibilities for investigations. Predictive policing could become a common practice in countries that have the technological capacities to use it. Palantir is on the rise.

And there are more things to worry about. How to deal with neurodata? Will old “lie detectors” be further developed into machines for “brain reading”? Mind-reading devices can now predict preconscious thoughts. AI-powered advances will threaten the privacy and autonomy of people in a new dimension.

In 2025, the UN Human Rights Council produced a first report about neurodata. Ana Brian Nougrères, the UN Special Rapporteur on the right to privacy, said: “Neurotechnologies are tools or devices that record or alter brain activity and generate neurodata that not only allow us to identify a person, but also provide an unprecedented depth of understanding of their individuality.” And she added: “Neurodata is highly sensitive personal data, as it is directly related to cognitive state and reflects unique personal experiences and emotions.” “Neurodata should be subject to the precautionary principle, enhanced accountability, and special measures to ensure security, confidentiality and limited circulation to prevent access or misuse, as well as manipulation, due to its potential to affect an individual’s mental integrity and thought processes negatively.” “While I welcome the potential mental health benefits of neurotechnologies, I am concerned that neurodata will not only allow access to what people think, but also manipulate people’s brains, leading to a violation of privacy in one’s own thoughts and decision-making,” Brian Nougrères said.

The UN Human Rights Council will have its 61st Session in Geneva in March 2026. Is this the right body to discuss this issue after the USA has left the Council? And where else should this issue be discussed? Nearly ten years ago, the first UN Special Rapporteur on Privacy in the Digital Age, Joe Cannataci, proposed a UN Convention against Mass Surveillance. But the efforts to draft such an instrument ended in the nowhere land. Will the same happen with the regulation of neurodata?

4. Artificial Intelligence

In the global discussion on artificial intelligence, we see a similar development to that in the field of cybersecurity. There are numerous efforts to discuss this global issue in an international context. But the major governments prefer a more nationalistic approach: my country first. And the global AI debate is increasingly overshadowed by the American-Chinese conflict over leadership in new AI applications.

4.1 The US-China AI Battle

The US is relying on its own strength. In July 2025, President Trump published his “AI Action Plan: Winning the Race”. For the US President, AI is “a national security imperative to achieve and maintain unquestioned and unchallenged global technological dominance. To secure our future, we must harness the full power of American innovation.” The US is now participating only in international projects to the extent that it serves its interests. This even applies to cooperation within the G7 (such as the OECD’s Global Partnership on Artificial Intelligence (GPAI) or the G7 Hiroshima AI Process/HAIP) and the ‘Bletchley Process’ initiated by the United Kingdom. The UN initiatives to form the International Scientific Panel on AI and the Global Dialogue on AI Governance are viewed with suspicion. The main opponent of the US is China.

China, on the other hand, is pursuing a more cooperative strategy, but one that leaves no doubt that China is striving for leadership in the AI race. China supports the UN initiatives on AI and has also advocated cooperation within regional associations of states such as APEC. In particular, it offers countries in the Global South cooperation in the development of AI capacities within the framework of its own “Global AI Governance Action Plan” (July 2025). In September 2025, China’s Vice Foreign Minister Ma Zhaoxu criticised “unilateralism and protectionism” in the UN Security Council, condemning “high fences around small spaces” designed to secure Western dominance and calling for internationally agreed AI governance models. At its core, however, China, like the US, is relying on its own strength and the innovative power of Chinese AI companies.

One of the new Chinese proposals is the establishment of a World Artificial Intelligence Cooperation Organisation (WAICO), headquartered in Shanghai. However, the details of how such a WAICO should be built remain unclear. The proposal, which was repeated by President Xi Jinping, Prime Minister Li Qiang and Foreign Minister Wang Yi at several high-level AI conferences, leaves open whether WAICO is to be an international organisation with governmental and non-governmental members or an NGO under the control of the Chinese party leadership.

4.2 The UN Initiative on Artificial Intelligence

In the UN, 2026 will see the start of the Independent International Scientific Panel on AI and the Global Dialogue on AI Governance.

A kick-off meeting of the Global Dialogue on AI Governance took place on the eve of the 80th UNGA in New York, chaired by UN Secretary-General Guterres. Guterres also published a ‘Call for Nomination’ to fill the AI panel. But the process has also raised concerns, not least because the US government made its opposition clear. At a meeting of the UN Security Council, Presidential Adviser Michail Kraitsos said: “We totally reject all efforts by international bodies to assert centralised control and global governance of AI. ... The path to this new AI world is found not in bureaucratic management, but in the freedom and duty of citizens, the prudence and cooperation of statesmen, and the independence and sovereignty of nations.’ He added: ‘The US is focused on establishing American AI as the global gold standard and enabling allies and trade partners to build their own sovereign AI ecosystems with secure American technology.”

However, the majority of UN members are positive about the two UN initiatives. Countries in the Global South, in particular, hope that they will provide impetus for overcoming the ‘digital divide’ and building an AI infrastructure. China is one of the main supporters. The EU countries also view both the AI panel and the AI dialogue positively. In contrast to the US government, some US companies such as Microsoft, Google and Meta have expressed their interest in both initiatives. Microsoft President Brad Smith has been campaigning for years to bridge the digital divide and encourages private investment in building AI infrastructure in the Global South.

Despite the rejection by the US, preparations for the first Global Dialogue on AI Governance are underway. It will take place together with the ITU’s AI for Good Summit and in parallel with the WSIS Forum in Geneva from 6 to July 10, 2026. However, the formation of the AI Panel has been delayed. The application deadline expired on October 31 2025. It is now up to UN Secretary-General Guterres to nominate the 40 members of the panel from the applications.

4.3 AI at G20, G7 and Bletchley

AI was also one of the main topics of the G20 summit on November 22, 2025, in Johannesburg. The “G20 South Africa Summit: Leaders’ Declaration” reaffirmed the G20 AI principles (2019), recognised the work of the G20 Task Force on Artificial Intelligence, Data Governance and Innovation for Sustainable Development, and launched a new ‘AI for Africa Initiative’ to serve as a voluntary platform for multilateral and multistakeholder cooperation. One should remember that the G20 did pioneer global AI policy in 2019 when it adopted the five AI policy principles developed by the OECD. The five principles—1. Inclusive growth, sustainable development and well-being, 2. Human-centred values and fairness, 3. Transparency and explainability, 4. Robustness, security and safety, and 5. Accountability—did set the standard for subsequent discussions on AI regulation, including the EU’s AI Act.

However, the absence of key G20 leaders at the summit in Johannesburg (US President Trump, Chinese President Xi and Russian President Putin did not attend the meeting) detracted from its success. In addition, the US behaved in a downright hostile manner towards the host country, South Africa and expressed scepticism about the role of the G20 in the field of AI.

This negative approach provoked a critical report by the Atlantic Council: “The US administration’s current approach to AI—marked by a preference for domestic industrial strategy and selective bilateral partnerships—reflects a hardening belief that multilateral governance is either futile or dangerous. In too many parts of Washington, there is a sense that global cooperation simply helps China; that multilateral institutions dilute US influence; and that if the US leads on innovation, it doesn’t need to lead on rules. This is a profound misreading of how power works in the digital age”.

AI was also the focus of the G7 Digital Ministers meeting in Montreal on December 9 2025. For years, the G7 has been a pioneer in the development of AI policy guidelines. The G7 Hiroshima AI Process (HAIP), launched in 2023, became a blueprint for policy frameworks for AI.

However, the “G7 2025 Industry, Digital and Technology Ministerial Declaration” focused less on the political aspects of AI development and more on practical ones. The preamble to the document generally refers to the G7 pursuing a ‘human-centric approach’ and supporting AI that drives innovation and growth and benefits people, mitigates negative externalities, promotes our economic and national security, respects applicable legal frameworks, including human rights, and is enabled through Data Free Flow with Trust (DFFT). In fact, however, the declaration is primarily concerned with accelerating innovative AI developments, especially in small and medium-sized enterprises (SMEs). An “SME AI Adoption Blueprint” and a “Toolkit for SMEs Deploying AI” were adopted. In addition, there will be an education initiative to ensure that the necessary personnel are available in the coming years. With a view to the future, a “G7 Joint Working Group on Quantum Technologies” was established.

But under Trump, the US sees its partners in the G7 less as a team for developing a uniform (Western) AI policy and more as a sales market for US tech companies. The ‘Golden Standard for AI’ is to be set by the US. The UN initiatives on AI or WSIS+20 were not on the agenda of the G7 digital ministers in Montreal.

The same US approach can be seen with regard to the so-called Bletchley Process, launched in 2022 by then British Prime Minister Sunak.

The first AI summit in Bletchley near London was attended by US Vice President Kamala Harris, French President Emanuel Macron and German Chancellor Scholz, as well as all the CEOs of the leading US tech companies and experts from China. The Bletchley Declaration recommended striking a balance between the opportunities and risks of AI development. It advocated for an “internationally inclusive network of scientific research on frontier AI safety that encompasses and complements existing and new multilateral, plurilateral and bilateral collaboration”.

At the follow-up meeting in Seoul in 2024, the establishment of a network of ‘AI Safety Institutes’ was agreed upon. At the third summit in Paris in February 2025, the US refused to approve the final document. The proposed regulatory measures would stifle innovation. Since then, the other initiatives of the Bletchley Process have also stalled.

India is now planning the 4th AI Summit for February 2026. It is the first AI summit in the Global South. According to Indian Prime Minister Modi, the aim of the New Delhi summit is to strengthen existing multilateral initiatives and set new priorities. The summit aims to set the course “to move from high-level political statements to demonstrable impact and tangible progress in global AI cooperation”. However, there is considerable scepticism that the New Delhi AI summit will go beyond nice words and produce practical progress for the Global South.

5. Management of Critical Internet Resources

Will the discussion on the management of critical Internet resources (CIR) come back in 2026? Will we see another attempt for a “governmental takeover”?

5.1 ITU

In November 2026, there will be the next ITU Plenipotentiary Conference (in Qatar). Some ITU member states could use this opportunity to challenge the existing mechanisms for the management of domain names, IP addresses and root servers.

In September 2025, Russia proposed in the ITU Council Working Group on Internet (CWG-Internet) to send a questionnaire to all ITU member states and ask them, inter alia, whether critical Internet resources should be managed by a private corporation. The CWG did not decide upon the questionnaire. But some ITU member states supported the idea.

Since 1998, when the ratification of the ITU led the AHC gTLD MOU to fail, and ICANN was established, some ITU member states tried again and again to get the management of CIRs under governmental control. They tried to use WSIS and failed in 2005. They tried to use the formula of “enhanced cooperation” and failed in the 2010s. Just recently, during the WSIS+20 review, they tried to funnel the idea into the IGF by establishing an independent “IGF Governmental Track”. But also this failed.

Will we see in Qatar in November 2026 the next round of this uphill battle? But before the Plenipot, there will be another CWG-Internet meeting in Geneva at the end of January 2026. And Russia has already tabled a proposal which should give “a significant and valuable response to actual challenges in the Internet ecosystem”. Russia sees the actual challenges as “the imperfection of the Internet governance system, and as a result, the fragmentation of the Internet, the privatization of the Internet by dominant digital platforms (big tech companies), issues of data transmission and protection, as well as challenges in the field of information security”.

5.2 ICANN

The best safeguard for the global Internet community, and in particular for ICANN, is to do an excellent job. If ICANN demonstrates that it is the right neutral steward of CIRs, there is no need for alternatives. Never change a winning team. The new gTLD program, which will start in 2026, is an excellent opportunity to demonstrate that the global Internet community can have no better steward than ICANN. ICANN has three meetings in 2026: Mumbai (March), Sevilla (June) and Oman (October).

This is also relevant for the management of IP addresses. The pressure of national governments to get a say with regard to IP address allocation is growing. IP addresses are the door opener for control and surveillance.

ICANN was, insofar as it was very wise in launching a debate on the review of its ICP-2 document, which sets the criteria for the establishment of new regional Internet registries (RIRs). This was, inter alia, triggered by the problems around AFRINIC.

ICANN’s response by organizing an action-oriented community dialogue, which included all concerned stakeholders, was the right one. ICANN’s mechanisms with the Address Supporting Organisation (ASO) and the Number Resource Organisation (NRO) are a good safeguard to keep the multistakeholder approach to IP address management resilient against efforts by some governments to play a bigger role.

There are certainly special local circumstances for all five RIRs. Africa (AFRINIC) is different from Europe (RIPE NCC), as Latin America (LACNIC) is different from the Asia Pacific (APNIC). However, there are universal procedures in place that guarantee that, within a multistakeholder process, all relevant stakeholders have a say if problems emerge. There are procedures in place for how ICANN and its Supporting Organisations (SOs) and Advisory Committees (ACs) cooperate with governments. There is the Governmental Advisory Committee (GAC), there are governmental liaisons in the SOs and ACs, and the SOs and ACs have their liaisons with the GAC. And if the ICANN Board violates the ICANN Bylaws, the empowered community, which includes governments via the GAC, can dissolve even the ICANN Board.

Certainly, the AFRINIC case is a stress test for the multistakeholder mechanism. The proposal for a Council of African Internet Governance (CAIGA) by Smart Africa, a partner of the Coalition for Digital Africa (CDA), has sparked discussion about a government takeover of IP address allocation in Africa. But so far, this stress test has demonstrated that the multistakeholder system is working. The discussions during the recent ICANN meeting in Dublin (November 2025) and the ongoing debates within the community, in particular within the ASO and the NRO, are demonstrating that the existing process can help to find sustainable solutions, also for AFRINIC. There is no need for “external governmental help.”

ICANN’s CEO Kurt Lidstrom has argued in his letter to ICANN’s Non-Commercial Stakeholder Group (NCSG): “Any change in the structure of the governance of an RIR must abide by the ICP-2 principles. As currently set out in the IG Blueprint, it is not clear to ICANN that the CAIGA proposal meets this test. If any proposal to modify AFRINIC’s governance structure were to come to ICANN, ICANN commits that it will neutrally and objectively evaluate that proposal. To be clear, ICANN has not pre-endorsed CAIGA or any other model as an acceptable model for the governance of an RIR”.

6. Outlook

2026 will be a complicated year for the governance of the Internet. It will be the first year with a permanent IGF. The modalities of the IGF have to be reviewed. Its leading bodies—the Multistakeholder Advisory Group (MAG) and the Leadership Panel (LP)—have to be repopulated. The UN Secretary-General will have to submit proposals for sustainable funding of the IGF by August 2026. And it needs an IGF Strategy for the next ten years if the IGF wants to remain “the primary multistakeholder platform for discussion of Internet governance issues” as it was recognized by the Global Digital Compact (GDC) in 2024.

The GDC itself will be reviewed in 2027. That means that the preparations for the review have to start already in 2026. As the WSIS+20 Outcome document has agreed, this should be done in close cooperation with other related processes, such as the WSIS Action Lines, the Sustainable Development Goals (SDGs) and others. This is also an opportunity that the various UN secretariats—ODET and UNDESA in New York, as well as UNGIS, ITU, IGF and UNCSTD in Geneva—can demonstrate their readiness to work hand in hand in organizing output-oriented multistakeholder cooperation. The intergovernmental UNGIS got a special mandate from the WSIS+20 Outcome document to facilitate such multistakeholder cooperation. Let’s wait and see how this will be organized in practice.

Probably the idea of an “Informal Multistakeholder Sounding Board” (IMSB), which was introduced as an experiment in the WSIS+20 review process to facilitate communication between governmental and non-governmental stakeholders, could be repeated for the GDC review or for other processes in the field of global digital policy making.

In the new political environment, this will not be an easy task. As Joana Kulesza has recently written in CircleD: “Formal multistakeholder institutions continue to operate and retain legitimacy. At the same time, authority over Internet-related policy outcomes is increasingly shaped through security alliances, criminal law frameworks, and enforcement-oriented cooperation mechanisms that function alongside, and often independently of, traditional Internet governance venues. This coexistence reflects a redistribution of power rather than a wholesale rupture.”

The problem is that it is nearly unavoidable that governments in the new political landscape will play a bigger role in Internet Governance. On the other hand, it is also unavoidable that non-state actors, such as the private sector, civil society and the technical and academic community, have to play a bigger role in the making of policy for Internet-related public policy issues. Non-governmental stakeholders will see a growing governmental role in Internet Governance as a threat to the multistakeholder model. Some governments will see a growing role of non-state actors as a threat to their control over their “national internet segment.”

The only way forward is to enhance understanding of the “respective roles and responsibilities” in digital policy-making and to move forward with innovative models of “shared decision making”, as agreed in the Tunis Agenda of 2005. The “Sao Paulo Multistakeholder Guidelines” (SPMG) from 2024 could be helpful to take the next steps. In the WSIS+20 Outcome Document, the 193 UN member States have “taken note” of the SPMGs. Now it is time to move from talking the talk to walking the walk.