n 2025, AI has exposed tensions at the heart of modern governance globally. Policymakers function through accumulation of data, consultation, negotiation and at a speed that is not compatible with the pace of modern technology. AI, most notably, reinvents itself before committees have finished agreeing on definitions.
For example, the EU AI Act was first proposed in 2021. The legislative process stretched across several years and by the time political agreement was reached, the landscape had changed with the arrival of technology such as OpenAI’s ChatGPT and Google’s Gemini, with AI–generated images that are almost impossible to distinguish from reality. Whole new sections had to be added to the AI Act while it was still being negotiated, and it still hasn’t been adopted.
The OECD has warned that governments need faster and more flexible ways to deal with technology as it is developing more quickly than existing monitoring structures. The UK’s Frontier AI Taskforce also argues that governance frameworks must adapt far quicker than in the past. This is backed by researchers at the Alan Turing Institute who have observed that the pace of AI development places pressure on the slower processes on which public institutions depend. Summits and papers have produced statements of shared concern, but each major jurisdiction has moved forward on its own path.
Cyberattacks once required sophisticated expertise and long preparation but this is no longer true. In the U.S., CISA has described how AI tools are helping even low-skilled attackers produce convincing scams and probe systems more effectively. Europol has warned that criminal and extremist groups are experimenting with large language models and the UK’s National Cyber Security Centre and the U.S. National Security Agency both state that AI is likely to make attacks faster and more adaptable.
Interestingly the issues faced at national, regional, and international levels and being faced by these large governmental and supranational bodies are often first identified and dealt with by informal and non–governmental networks. For example, the Shadowserver Foundation and Citizen Lab have repeatedly identified emerging threats before official bodies have done so because they can move more quickly and operate across borders without procedural delay.
Whilst not regulating technology or building it, alliances are looking at some of the questions around the development of AI. Organizations such as the Ada Lovelace Institute, the Leverhulme Centre for the Future of Intelligence, the Montreal AI Ethics community, All Tech is Human and my own think tank, Saviesa.
What we have seen in 2025 is clear. Informal power is not only increasing, it is becoming essential to effective governance. These networks can see the future forming while institutions are still anchored in the present. The work of these networks does not replace the authority of the state, but it complements it, strengthens it and, in many cases, precedes it. In a world shaped by accelerating technology, informal power has become an indispensable part of the governance ecosystem.
a global affairs media network
The silent efficiency of informal power
Photo by Luke Jones on Unsplash.
December 11, 2025
AI is moving faster than governments can regulate, making informal networks essential early responders in today’s rapidly shifting governance landscape, writes Leonor Diaz Alcantara.
I
n 2025, AI has exposed tensions at the heart of modern governance globally. Policymakers function through accumulation of data, consultation, negotiation and at a speed that is not compatible with the pace of modern technology. AI, most notably, reinvents itself before committees have finished agreeing on definitions.
For example, the EU AI Act was first proposed in 2021. The legislative process stretched across several years and by the time political agreement was reached, the landscape had changed with the arrival of technology such as OpenAI’s ChatGPT and Google’s Gemini, with AI–generated images that are almost impossible to distinguish from reality. Whole new sections had to be added to the AI Act while it was still being negotiated, and it still hasn’t been adopted.
The OECD has warned that governments need faster and more flexible ways to deal with technology as it is developing more quickly than existing monitoring structures. The UK’s Frontier AI Taskforce also argues that governance frameworks must adapt far quicker than in the past. This is backed by researchers at the Alan Turing Institute who have observed that the pace of AI development places pressure on the slower processes on which public institutions depend. Summits and papers have produced statements of shared concern, but each major jurisdiction has moved forward on its own path.
Cyberattacks once required sophisticated expertise and long preparation but this is no longer true. In the U.S., CISA has described how AI tools are helping even low-skilled attackers produce convincing scams and probe systems more effectively. Europol has warned that criminal and extremist groups are experimenting with large language models and the UK’s National Cyber Security Centre and the U.S. National Security Agency both state that AI is likely to make attacks faster and more adaptable.
Interestingly the issues faced at national, regional, and international levels and being faced by these large governmental and supranational bodies are often first identified and dealt with by informal and non–governmental networks. For example, the Shadowserver Foundation and Citizen Lab have repeatedly identified emerging threats before official bodies have done so because they can move more quickly and operate across borders without procedural delay.
Whilst not regulating technology or building it, alliances are looking at some of the questions around the development of AI. Organizations such as the Ada Lovelace Institute, the Leverhulme Centre for the Future of Intelligence, the Montreal AI Ethics community, All Tech is Human and my own think tank, Saviesa.
What we have seen in 2025 is clear. Informal power is not only increasing, it is becoming essential to effective governance. These networks can see the future forming while institutions are still anchored in the present. The work of these networks does not replace the authority of the state, but it complements it, strengthens it and, in many cases, precedes it. In a world shaped by accelerating technology, informal power has become an indispensable part of the governance ecosystem.
