Artificial Intelligence is rapidly becoming part of the infrastructure of our societies. It influences how we access information, how organisations make decisions, how people learn and work and how governments provide services.
Increasingly, it also shapes how we understand the world around us. Much of the current debate focuses on the risks of AI. There are good reasons for this. AI can amplify misinformation and political polarisation, enable new forms of surveillance, reproduce discrimination and concentrate economic and political power. And its military applications raise even more fundamental questions.
We need to address these risks for sure. In addition there is a second question that deserves much more attention.
What if we looked at AI not only from the perspective of preventing harm, but asked how it could contribute to peace?
This is where the concept of Positive Peace becomes interesting.
From the absence of violence to Positive Peace
The Institute for Economics & Peace describes Positive Peace as the attitudes, institutions and structures that create and sustain peaceful societies.
The idea behind it is straightforward: peace is more than the absence of war or direct violence. It depends on the conditions under which people live together. It raises questions such as:
- Do institutions work in favour of citizens and do people trust them?
- How are resources and opportunities distributed fairly?
- Can people access information and are their rights protected?
- Do people have the skills and opportunities to shape their own lives?
The original Positive Peace framework describes eight interconnected pillars, ranging from well-functioning government and equitable distribution of resources to human capital, information and good relations with neighbours. AI is beginning to affect almost every one of these conditions.
That is why the relationship between AI and peace goes much further than using an AI tool in mediation or conflict analysis. AI itself is becoming part of the environment in which cooperation, inequality, trust and conflict develop.
So I tried to translate the logic of Positive Peace into the world of AI. The result is not intended as a definitive framework. It is a proposition for discussion: eight conditions that could help AI contribute to Positive Peace.
01Human-Centred and Accountable AI Governance
The first pillar is about power. AI systems increasingly influence decisions in employment, education, healthcare, finance and public administration. As their role grows, so does a fairly basic question: Who gets to decide what these systems are designed to do?
At the moment, much of that power sits with technology companies, developers and governments. But people who are affected by AI systems need meaningful ways to understand and challenge their decisions.
Human-centred AI governance therefore requires transparency, accountability and democratic oversight. It also requires us to preserve something more fundamental: human agency. People should not gradually become the objects of systems they no longer understand and cannot influence.
This makes AI governance more than a technical or regulatory issue. It is ultimately a question about the distribution of power.
02Equitable Access and Distribution of AI Benefits
AI has the potential to generate enormous economic and social value. And there is no guarantee that this value will be distributed broadly. Much of the infrastructure behind advanced AI is already highly concentrated: computing power, capital, data, expertise and foundation models.
This raises a question I believe will become increasingly important: Who receives the AI dividend?
If AI dramatically increases productivity while most of the economic gains accumulate among a relatively small number of corporations, regions and individuals, it could deepen existing inequalities. And the issue is not solved by giving everybody access to ChatGPT or another AI assistant. People, communities, small organisations, cities and regions need the capacity to use AI to create value themselves.
For me, this is where the discussion around AI connects directly with ideas such as Community Wealth Building. AI should not simply extract data, knowledge and value from communities. It should help them create and retain value locally. That would make the distribution of AI capabilities an important element of Positive Peace — and it is one reason why we currently work on our new product, the Community Action Platform.
03A Safe and Just Digital Environment
There is no Positive Peace without safety. People need protection from algorithmic discrimination, manipulation, fraud, cyber abuse and intrusive surveillance. Military applications of AI and autonomous weapons raise an even more fundamental set of questions.
But there is a tension here. If we respond to the dangers of AI with ever greater surveillance and control, we may solve one problem by creating another. A safe digital environment therefore also needs privacy, rights and limits on power.
Human rights and human agency are not obstacles to AI safety. They are part of it.
04Free Flow of Knowledge and Information
This is perhaps one of the areas where AI's potential is easiest to see. Generative AI can give people access to knowledge that was previously difficult or expensive to obtain. A student can have something resembling a personal tutor. Language barriers can be reduced. Small organisations can analyse information that once required specialist teams. That could be enormously empowering.
At the same time, the same technology makes misinformation and synthetic content cheaper and easier to produce. Personalised persuasion can become much more sophisticated — and by that, invisible.
The challenge is therefore not simply access to information but the ability to navigate an increasingly complex information environment — instead of further creating AI systems that decide for us what is true or what we should believe. A much more resilient approach is to strengthen people's capacity to question information, compare perspectives and form their own judgement.
Critical thinking may become an increasingly important part of peace infrastructure.
05Sustainable and Responsible Use of Resources
AI looks digital, but its infrastructure is physical. Data centres need electricity and water. Chips require minerals, complex supply chains and large industrial facilities. The rapid expansion of AI therefore has ecological as well as geopolitical consequences.
Some regions facing the high water and electricity demands of newly built data centres have already begun to address these problems. In Uruguay, an earlier design for a Google data centre was projected to use up to 7.6 million litres of water per day for cooling. The controversy unfolded during the country's worst drought in decades and triggered significant public opposition. The project was subsequently redesigned with a different cooling approach.
The case illustrates a broader Positive Peace question: when AI infrastructure competes for scarce resources, who participates in deciding how those resources are allocated?
This topic needs to become part of a much bigger conversation about AI and peace — before data centres are even planned. At the same time, AI can help us manage energy systems, understand environmental change, improve resource use and support regenerative forms of economic development. The challenge is to do both: reduce the environmental costs of AI itself while using its capabilities to address environmental problems.
A technology cannot contribute to long-term peace while undermining the ecological systems on which societies depend.
06Strong Social Cohesion and Trust
This is the pillar I find particularly interesting. Digital technologies increasingly mediate relationships between people. Algorithms influence what we see, whose opinions we encounter and which stories receive our attention.
Much of today's digital infrastructure has been designed around engagement. And outrage happens to generate engagement remarkably well. What would happen if we started designing some AI systems around a different objective: the objective of understanding?
AI could help people explore perspectives they disagree with. It could help identify misunderstandings before they escalate. It could support communication across languages and cultures. For mediation and peacebuilding, there are very concrete possibilities people are already working on: AI can help analyse conflicts, structure large amounts of information, identify underlying interests or help people prepare for difficult conversations.
This does not mean outsourcing conflict resolution to machines. I would argue almost the opposite.
AI should help humans become better at conflict. Better at listening. Better at seeing different perspectives. Better at identifying common interests. Better at finding solutions together.
That would be a very different objective from simply maximising engagement.
07Empowered People and Human Capability
There is another possible trajectory of AI that worries me. If AI writes for us, searches for us, navigates for us and increasingly recommends decisions for us, convenience can slowly turn into dependency. But Positive Peace requires something different.
AI should expand human capabilities rather than quietly replace them. That requires education, AI literacy and critical thinking. But it also requires something broader: agency and accountability.
We often ask what AI will be able to do for humans. Perhaps we should ask more often: What will humans become able to do because they have AI?
For me, that is an important distinction. The aim should not be humans stepping out of the process. It should be humans becoming more capable of understanding complexity and acting within it.
08Ethical and Purpose-Driven Innovation
A great deal of technological development is driven by two rather simple questions:
- Can we build it?
- Can we make money from it?
Both are legitimate questions. Neither tells us much about the society that emerges from the technology. So I propose to ask a third question: What is this technology actually for?
Purpose-driven AI, or AI for wellbeing, does not mean being against innovation or commercial success. It means taking a wider view of what successful innovation looks like. Human dignity matters. Social cohesion matters. Sustainability matters. The ability of people to shape their own lives matters — and of course, peace matters.
Perhaps some of the most valuable AI applications will not be those that replace the greatest number of human tasks, but those that enable people to solve problems together that they previously struggled to solve.
THE PROPOSITION
From AI Alignment to Peace Alignment
This leads me to what I think is one of the more interesting implications of the framework. At Peace & AI, we talk a lot about AI alignment: how to ensure that AI systems act according to human intentions and values. But alignment alone does not answer the societal question.
Imagine an AI system that is extremely good at doing exactly what its user wants. That user might want to manipulate consumers, maximise outrage, spread propaganda or exploit people's vulnerabilities. The AI could be perfectly aligned with its user and still contribute to a less peaceful society.
We therefore need to broaden the idea of alignment. One way of thinking about this could be Peace Alignment — a concept we will explore in depth in a dedicated essay.
I don't mean programming a particular political worldview into AI. Nor should AI decide what peace looks like for us. The idea is almost the reverse: a peace-aligned AI should strengthen people's ability to make decisions themselves.
- Agencyrather than dependency
- Understandingrather than manipulation
- Dialoguerather than polarisation
- Cooperationacross differences
AI should not become an arbiter of peace. It should help us become better at creating peace ourselves.
What kind of society do we want to build with AI?
The debate about Artificial Intelligence is often presented as a series of binary choices: innovation or regulation, humans or machines, technological progress or ethics. I don't think these are the most useful questions. AI is already becoming part of our societies. The more important question is what kind of relationship we build with it.
- Will AI concentrate capabilities or distribute them?
- Will it optimise mainly for attention or help us understand one another?
- Will it make people more dependent or more capable?
- Will it automate judgement or help us make better judgements?
- Will we primarily use AI to compete more effectively — or can it also help us cooperate more intelligently?
Technology does not answer these questions for us. We as humankind have to step up and find the answers.
The question is not only how intelligent AI will become. The question is what kind of society we become alongside it.
AI shapes us. We shape AI. And together, we can shape AI for peace.