Originally published on the PeaceTech Alliance Blog. Based on the panel discussion “Do No Harm: Who Decides What Technology Peace Needs?” at The PeaceTech We Want Summit 2026, 19 October 2026, organised by the PeaceTech Alliance and the Open Knowledge Foundation. Contributions have been paraphrased from the session transcript for clarity. The proposed peacebuilder-centred design approach was discussed as a possible direction for future collective work, not as an established framework.
Technology is increasingly presented as a way to address some of the world’s most difficult peace and conflict challenges. Artificial intelligence can help analyse online narratives, digital platforms can support dialogue between divided communities, and data-driven tools can assist peacebuilders in understanding emerging risks. Yet there is a fundamental question that is too often overlooked in the development of these technologies: who gets to decide what the problem is, and who determines what a successful solution should look like?
This question formed the basis of Do No Harm: Who Decides What Technology Peace Needs?, a panel at The PeaceTech We Want Summit 2026, organised by the PeaceTech Alliance and the Open Knowledge Foundation on 19 October 2026. Moderated by Nathan Coyle, Principal of the PeaceTech Alliance, the discussion brought together Benjamin Cerigo, Peace Tech Architect at Build Up, Lena Slachmuijlder, Senior Advisor of Digital Peacebuilding at Search for Common Ground and Co-Chair of the Council on Tech and Social Cohesion, and Naomi Kilungu, Founder and Executive Director of AI for Peace Africa.
Since its formation in 2025, the PeaceTech Alliance has championed the integration of Do No Harm and community development into the design of PeaceTech. The emphasis has been on moving beyond simply developing technologies for peacebuilders, towards designing with communities, recognising local expertise and ensuring that projects leave behind lasting benefits. The panel offered an opportunity to explore these principles more deeply, bringing together different perspectives on how they can be translated into practice and where existing approaches might need to go further.
Although the discussion examined the opportunities presented by artificial intelligence and other emerging technologies, its central concern was not technological capability. Instead, the panel explored the relationship between technology, agency and community development, questioning whether established approaches to human-centred design go far enough when technologies are being introduced into communities experiencing conflict.
The conversation also highlighted an important distinction between developing technology for peace and developing technology with those who are working to achieve it. While the two approaches may appear similar, they can produce very different outcomes, particularly when the people designing and funding technologies are geographically, culturally or institutionally removed from the communities expected to use them.

The cycle the panel described, and three questions to break it.
Agency must mean more than participation
Lena Slachmuijlder opened the discussion by examining the meaning of agency in peacebuilding. Drawing on her experience at Search for Common Ground, she described agency as the belief that individuals possess the skills, ability and recognition necessary to influence their circumstances. In environments affected by conflict, where people may feel powerless to change the conditions around them, developing this sense of agency can be an important outcome in its own right.
Technology can contribute to that process. Slachmuijlder pointed to emerging research and practical examples in which AI-assisted dialogue tools help individuals prepare for difficult conversations or participate more confidently in deliberative processes. The opportunity to explore an argument, rehearse a conversation or make mistakes privately can give people greater confidence when engaging with others.
However, she also distinguished between technology that strengthens an individual’s agency and the much larger question of who governs the technology itself. People may benefit from using AI tools without having any meaningful influence over how those tools are developed, what information they use or the decisions made by the companies that control them.
This creates a difficult relationship between empowerment and dependency. A technology may provide immediate benefits while leaving its users dependent on systems whose priorities, governance and future development remain outside their control.
For peacebuilding, the distinction is particularly significant. Participation in a digital activity does not necessarily mean participation in the decisions that shape it. Communities may be consulted, invited to test a product or encouraged to provide feedback, yet still have little influence over the original problem definition, the ownership of the data or whether the technology continues to exist once external funding ends.
Agency, therefore, cannot be measured simply by the number of people who use a tool or attend a consultation. It must also consider whether people have the ability to influence the decisions that affect them.
Community development is not an optional extra
For Naomi Kilungu, the question of agency begins much earlier, with the definition of the problem itself.
Drawing on her work with AI for Peace Africa, Kilungu argued that technologies are too frequently developed outside African communities and subsequently introduced as solutions to problems that have not been adequately understood. This approach risks overlooking the economic, political, cultural and social conditions that shape conflict, while assuming that a technology designed in one environment can be transferred into another with limited adaptation.
She emphasised that peace cannot be separated from community development. Economic opportunity, resource management, governance, education and the legacies of conflict all influence how peace is understood and experienced. Technologies intended to support peacebuilding must therefore be designed with an understanding of these wider conditions, rather than treating conflict as a technical problem that can be isolated from its social environment.
The first responsibility of a technology project, Kilungu argued, is to establish what the community actually needs. This requires meaningful involvement from people experiencing the problem, including those whose perspectives may not be represented through conventional consultation processes.
Local language and culture are equally important. AI systems trained primarily on widely available datasets may struggle to interpret the language, expressions and social context of particular communities. In peacebuilding, where the meaning of a phrase or the interpretation of a political message can depend heavily on local circumstances, these limitations can have serious consequences.
Kilungu described work being undertaken by AI for Peace Africa ahead of Kenya’s 2027 elections, involving the development of tools to better understand language and political communication in specific communities. Rather than attempting to apply one model across the country, the approach considers differences in language, culture and local conflict dynamics.
The example illustrates why community involvement must extend beyond providing data. Local experts are needed to interpret information, understand its significance and determine whether the conclusions produced by a system are appropriate. Their involvement is also important in establishing how a technology will be maintained, governed and used over time.
The question of data ownership runs throughout this process. Kilungu raised concerns about the collection of African data without sufficient local knowledge or control over how that information is subsequently used. When communities contribute information to a project but have little influence over its governance, they may become sources of data rather than partners in development.
This is not simply a matter of representation. It affects the practical usefulness, legitimacy and sustainability of the technology itself.
The panel also challenged the assumption that communities in the Global South lack the technical ability to develop their own solutions. Kilungu described an expanding ecosystem of African innovators working on technologies for democratic accountability, violence prevention and peacebuilding. The difficulty, she argued, is often not the absence of ideas or expertise, but the lack of support, collaboration and opportunities to develop existing initiatives into sustainable projects.
This suggests a different role for international organisations and technology partners. Rather than arriving with predetermined solutions, they could invest in strengthening local capabilities, supporting existing innovators and creating partnerships in which communities can negotiate the terms of collaboration, including data governance and ownership.
Designing with communities, not simply for them
Benjamin Cerigo brought the perspective of a software engineer working directly on peace technology. As a Peace Tech Architect at Build Up, he has been involved in developing digital tools intended to support peacebuilding practitioners, including the organisation’s Phoenix social media analysis platform.
His contribution demonstrated how the principles of community-led development can influence practical engineering decisions.
Cerigo described Build Up’s earlier work with local organisations in Syria through its Digital Steps initiative, where participating organisations were supported in identifying digital initiatives that responded to their own peacebuilding priorities. The approach placed local needs at the beginning of the development process, rather than introducing a predetermined technical solution.
He also reflected on the development of Phoenix and the transition from a more traditional approach, in which extensive features are planned in advance, towards an iterative process that tests functionality against the needs of practitioners.
One example concerned the classification of social media information. More sophisticated techniques, including sentiment analysis, initially appeared attractive. However, the limited availability of appropriate language datasets and the practical requirements of peacebuilders led to a simpler approach based on text matching.
Although less technically elaborate, this approach could be applied across languages without requiring the development of specialised classification models for each one. It offered a more practical response to the actual needs of users.
The example demonstrates a problem that extends beyond peace technology. Innovation is often assessed according to technical sophistication, even when a simpler approach would be more accessible, adaptable or useful.
Cerigo also acknowledged the assumptions that engineers bring into development processes. Software engineering is built around identifying problems and constructing solutions, but the social and political conditions surrounding peacebuilding cannot always be translated into technical requirements.
The availability of increasingly powerful AI development tools may intensify this tendency. As building software becomes easier, there is a risk that the ability to create a tool is mistaken for evidence that the tool is needed.
Slachmuijlder reinforced this point by describing how the ability to build monitoring tools through natural-language coding can create an illusion that collecting and presenting information will automatically lead to decisions or cooperation. In environments where trust is weak, however, information alone cannot resolve the relationships and disagreements that prevent people from acting together.
The lesson is not that peacebuilders should avoid technology, or that technical innovation is inherently problematic. Rather, technology development must remain connected to the social processes through which peace is built.
Should peacebuilding have its own design approach?
Building on these perspectives, the panel considered whether conventional human-centred design provides a sufficient framework for peace technology.
Human-centred design has become an established approach to developing products and services around the needs, experiences and capabilities of users. However, peacebuilding introduces additional considerations that may not be adequately captured by conventional usability or accessibility assessments.
Coyle introduced the possibility of a complementary peacebuilder-centred design approach, one that would account for the environments in which peace practitioners operate, the relationships they manage and the particular responsibilities associated with working in conflict-affected communities.
Importantly, this was presented as a question for collective consideration rather than a completed framework. The intention was not to replace human-centred design, but to explore whether additional considerations could help technology developers better understand the circumstances in which peacebuilding takes place.
Even identifying the intended users can be complicated. Individuals involved in peacebuilding may describe themselves as community workers, activists, humanitarian practitioners or members of local organisations rather than peacebuilders. A design process that consults only formally recognised peacebuilding institutions could therefore overlook people undertaking essential work within communities.
The panel’s responses highlighted several considerations that such an approach would need to address, including local legitimacy, cultural understanding, trust, community participation, data governance and the ability of people to maintain control over technologies after their initial development.
Kilungu emphasised the importance of collaboration between local innovators, peacebuilders, governments and international organisations, while ensuring that communities retain the ability to define their priorities and negotiate the conditions of technological partnerships.
Slachmuijlder argued that technology should also be understood in relation to the wider purpose of peacebuilding. Beyond documenting harms or identifying malicious activity, digital tools can help people develop trust, collaborate across differences and participate in collective decisions.
She pointed to deliberative technologies, including examples from Taiwan, as illustrations of how digital participation can contribute to stronger relationships between citizens and decision-makers when it is connected to meaningful processes of public decision-making.
Cerigo offered a caution about the belief that peace itself can be engineered. Technical systems can support social processes, but they cannot substitute for the human relationships, judgement and negotiation through which those processes develop.
Together, these perspectives suggest that peacebuilder-centred design would need to consider more than whether a technology functions correctly or meets the immediate requirements of its users. It would also need to examine how the technology influences relationships, distributes power and affects the long-term capabilities of the communities involved.
A central consideration is whether community development should itself become a measure of successful design.
If a technology performs its intended function but leaves a community no better equipped to address future challenges, can it reasonably be considered successful? Conversely, a relatively simple tool that strengthens local knowledge, builds technical capacity and supports independent decision-making may create benefits that extend well beyond its immediate use.
This shifts the assessment of peace technology from what a product delivers to what remains with the community after the project has ended.
Do No Harm must begin before development
The final part of the discussion returned to the principle of Do No Harm, a concept familiar across peacebuilding, humanitarian assistance and international development.
The principle is commonly associated with the work of Mary B. Anderson and the recognition that external interventions can unintentionally reinforce divisions, redistribute resources in harmful ways or alter relationships within conflict-affected societies.
Applied to technology, Do No Harm provides a means of examining the consequences of digital interventions beyond their intended benefits. A system designed to identify misinformation, for example, may produce useful information while also creating risks relating to surveillance, political exclusion, data misuse or the misinterpretation of local language.
For Coyle, Do No Harm also offers a shared language through which peacebuilders, technologists and policymakers can discuss responsible development. However, the panel recognised that different professional communities may interpret the principle differently.
Cerigo suggested that policymakers may approach Do No Harm through compliance, while peacebuilders may focus on the relationships, dividers and connectors within a community. Technology developers, meanwhile, may interpret it through technical risk management.
These perspectives are not necessarily incompatible, but they can produce different assessments of what constitutes harm and who is responsible for identifying it.
Cerigo illustrated this challenge through Build Up’s experience with electoral work in Puntland, where local participants identified value in a technical approach that external actors did not regard as a valid way of assessing democracy. The example raised questions about whose judgement is treated as legitimate when external technical standards conflict with local preferences.
Do No Harm, therefore, cannot be reduced to a checklist applied by external experts. It must also examine who has the authority to define harm, whose experience is considered credible and how communities can influence decisions when disagreements arise.
Kilungu similarly highlighted the importance of representation in technology governance. While stronger governance can support the development of responsible technologies, the processes through which rules and standards are established are not automatically neutral. Decisions about who participates in governance can reproduce existing inequalities and exclude perspectives from communities most affected by technological interventions.
Slachmuijlder argued that Do No Harm is particularly valuable because AI and digital technologies are being adopted rapidly across sectors that extend well beyond peacebuilding, including education, healthcare and public administration.
These systems may achieve their stated objectives while simultaneously creating new inequalities or reinforcing barriers to participation. A digital public service may improve efficiency, for example, while excluding people who lack connectivity, digital skills or trust in the institutions providing it.
Peacebuilders have experience examining these unintended consequences through their work in humanitarian assistance, development and conflict transformation. Slachmuijlder suggested that this experience provides a foundation for bringing Do No Harm into wider digital transformation processes.
The implication is that peacebuilding expertise should not be confined to technologies explicitly labelled as PeaceTech. It can also contribute to the responsible design and deployment of technologies across other sectors, particularly where digital interventions affect vulnerable populations or existing social divisions.
Yet the panel’s discussion also exposed a limitation of approaches focused exclusively on avoiding harm.
A technology might successfully minimise identifiable risks without contributing meaningfully to community development, strengthening local capabilities or addressing the conditions that make communities vulnerable in the first place.
This is where Do No Harm and community development need to be considered together.
Avoiding unintended harm is essential, but it should not become the only measure of responsible technology. A more comprehensive approach would also ask whether communities have gained knowledge, ownership, resources and greater influence over the decisions that affect them.
Such considerations are particularly important for projects supported through short-term funding arrangements. External organisations may develop a tool, collect data, publish their findings and move on, leaving communities with limited ability to maintain the technology or benefit from the information they helped produce.
The consequences can extend beyond the failure of an individual project. Repeated experiences of external intervention without lasting benefits can undermine trust and make future collaboration more difficult.
Community development offers a way of addressing this problem by treating local capability and ownership as part of the intended outcome from the beginning. It requires developers and funders to consider not only whether a technology can be delivered, but also who will govern it, who will maintain it and whether the community has the resources and authority to continue using or adapting it independently.
This does not mean that every peace technology project must produce a permanent product or establish a new local institution. It does mean that decisions about sustainability, ownership and the eventual conclusion of a project should be made with communities rather than imposed upon them.
From principles to practice
The discussion at The PeaceTech We Want Summit did not produce a new design standard or a definitive framework for responsible PeaceTech. Nor was that its purpose. Instead, it brought together perspectives from peacebuilding practice, African technology development, software engineering and digital governance to examine what existing approaches may be missing.
Several points of agreement emerged. Technology should begin with an understanding of the problem rather than a predetermined solution. Communities should have meaningful influence over the technologies intended to serve them. Local expertise, culture and language must shape development decisions, while governance and ownership should be considered throughout the process rather than after deployment.
The panel also demonstrated that technical sophistication is not necessarily a measure of usefulness. Simpler, more accessible technologies may be better suited to the circumstances of peacebuilders, particularly where infrastructure, language resources or long-term technical support are limited.
Perhaps most importantly, the conversation challenged the tendency to treat community participation as one stage within a technology project. Community development is a continuing process, involving relationships, trust, skills, institutions and the ability of people to make decisions about their own futures.
For PeaceTech, this raises a more demanding standard of success. A technology should not be assessed solely according to whether it functions as intended, reaches its target users or produces measurable outputs. Its effects on community agency, ownership and long-term capability also deserve attention.
Do No Harm provides an important foundation for this work, particularly in recognising the unintended consequences that technological interventions can create. However, its value depends on how it is applied and who participates in determining what harm means.
The central question is therefore not simply whether technology can contribute to peace. It is whether the processes through which that technology is designed, governed and introduced strengthen the people and communities whose peace it is intended to support.
As the development of AI and other digital technologies becomes increasingly accessible, the temptation to build solutions will only grow. The challenge for PeaceTech is to ensure that the ability to build something does not replace the responsibility to understand whether it is needed, who should shape it and what it will leave behind.
PeaceTech should not simply be technology designed to do no harm. It should be technology developed with communities in ways that strengthen their ability to build and sustain peace themselves.








You must be logged in to post a comment.