Last week, it combined three major events devoted to the governance, practical uses and developmental impact of artificial intelligence. Taken together, the first United Nations Global Dialogue on AI Governance, the AI for Good Global Summit and the World Summit on the Information Society Forum revealed something more important than the usual international declarations. They exposed the widening distance between technological change and the ability of governments, international organizations and societies to control its consequences.
There is no shortage of references to ethics, transparency, inclusion, human oversight and responsible innovation. The real problem is that the institutions capable of enforcing these principles remain weak, fragmented or absent. The companies developing the most powerful systems move across borders with considerable freedom, while national regulators remain limited by jurisdiction, technical capacity and political pressure.
This imbalance was especially visible in the United Nations dialogue. For the first time, AI governance was discussed in a universal setting in which developing countries had a formal place. Yet representation is not the same as influence. Most countries remain dependent on technologies, cloud systems, data centers and models designed elsewhere. They may participate in writing the language of global governance while remaining unable to shape the infrastructure on which it depends. The global AI divide is therefore deeper than the difference between those who can use an AI application and those who cannot. It is also a divide in computing power, data, finance, research, language and the ability to set standards. Unless this imbalance is addressed, many developing countries will remain consumers of systems created elsewhere, according to priorities that do not necessarily reflect their own societies.
The AI for Good Summit showed the other side of the picture. It presented useful applications in health, education, agriculture, disaster management and accessibility. These examples demonstrated that AI can improve public services and help solve real problems. But the phrase "AI for good" can also become too comfortable. The same cloud systems that support hospitals and universities can support surveillance, intelligence analysis and military targeting. Technology does not choose its purpose. Governments and companies do.
This contradiction came to the surface when activists interrupted a presentation by Amazon's chief technology officer to protest Project Nimbus, the cloud contract awarded to Amazon and Google by the Israeli government. Nimbus illustrates how a civilian cloud contract can become part of a broader military and intelligence infrastructure, even when the precise operational uses of that infrastructure remain shielded from public scrutiny. The protesters accused the companies of providing infrastructure that supports Israeli military operations against Palestinians. Publicly available evidence show that commercial technology companies are becoming deeply embedded in military and intelligence systems through arrangements whose full details remain hidden.
Google employees were dismissed after protesting against Project Nimbus. But the controversy is not limited to Amazon and Google. Microsoft employees also challenged their company's cooperation with the Israeli military through Microsoft’s Azure platform. Investigations by major international news organizations found that commercial cloud and AI tools were used in intelligence analysis and in processes connected to target selection in Gaza and Lebanon. Microsoft later restricted some services after concerns emerged over the storage of surveillance data relating to Palestinians.
The central issue is not whether a server, database or AI model is itself a weapon. Modern military action increasingly depends on a chain that begins with data collection, passes through cloud storage and algorithmic analysis, and may end with a human decision to strike. Responsibility is divided among several actors, each claiming to control only one stage. The final act may be visible, but the chain that produced it remains obscure. This is close to the concern I raised earlier in my last article about the renewed attention to the series Person of Interest. The problem is not simply that machines are becoming more capable; it is that power is being distributed through systems that make responsibility harder to identify and easier to avoid.
WSIS added another dimension. The digital divide can no longer be reduced to which people are more connected and which are less or not connected. It now includes affordability, digital skills, computing power, language, data, cybersecurity and the ability to produce technology rather than merely consume it. Here again, the gap between declaration and implementation remains wide. Inclusion without capacity may simply reproduce dependence in a more modern form.
For Egypt, which chaired the 2026 WSIS Forum, the opportunity lies not in the visibility of the chairmanship alone, but in drawing the right conclusions. Egypt should link its AI strategy, cybersecurity policy, digital-government programs and education plans within one coherent framework. Technology contracts should be treated as matters of sovereignty, not routine procurement. Egypt must know where its data are stored, how they are used, who can access them and under what conditions services can be suspended. Public bodies using AI in health, education, welfare or security should be subject to clear rules and genuine human responsibility.
Egypt should also work with Arab and African partners to strengthen Arabic and local languages, computing capacity and negotiating power. The objective should not be technological isolation, but the ability to choose, regulate and negotiate from a stronger position.
Geneva did not produce a global constitution for AI. It did something more modest, but useful. It clarified the central political question. The future of AI will not be decided only by what machines can do, but by who controls them, who benefits from them and who is held responsible when they cause harm. That is the debate Egypt should now help lead.
Short link: