By JoAnne Wadsworth, Communications Consultant for the G20 Interfaith Forum
———
On July 16, 2026, the G20 Interfaith Forum held a webinar entitled “AI and Peacebuilding” as part of its “AI and Human Flourishing” series. Speakers included John, Lord Alderdice, a psychiatrist and member of the House of Lords who, as leader of the Alliance Party, was a key negotiator of the 1998 Good Friday Agreement and went on to serve as the first Speaker of the Northern Ireland Assembly; he holds honorary academic appointments at Oxford and is founding chair of the Concord Foundation. Also speaking was Julian Weinberg, founder and executive director of Dialogue Action, who has more than fifteen years’ experience leading high-level political dialogue processes across Europe, Southeast Asia, North Africa, and Western Asia, is a member of the Computational Humanities Lab at the University of Oxford, and is a non-resident fellow at Stanford’s Center on International Conflict and Negotiation. Pieter François, Professor of Cultural Evolution at the University of Oxford, Director of the Computational Humanities Lab, and co-chair of the G20 Interfaith Forum’s subgroup on Freedom of Religion or Belief, moderated and took part in the discussion.
Julian Weinberg
Weinberg opened by tracing how his interest in the subject began. Nearly three years ago, as the wider world was watching the arrival of generative AI, a board member asked him what its impact on mediation and peacebuilding would be. His first instinct was skepticism—peacebuilding, he reasoned, is fundamentally about people looking at one another across a table and developing relationships. But as he and his colleagues thought it through, they concluded that AI was certain to affect their field as it was affecting so many others, and that it deserved serious attention.
To pursue the question, Dialogue Action convened a two-year series of roundtables—one hosted at the Alan Turing Institute and another at Stanford—that brought technologists, policymakers, and peacebuilders into the same room. Two working questions guided the effort: which challenges in mediation, conflict prevention, and peacebuilding might emerging technologies help with, and what practitioners and technologists might build or attempt if each understood the other’s world better. The roundtables produced a white paper landscaping the assessment as a contribution to a rapidly developing area.
Weinberg reported that some organizations, including a UN innovation unit within the DPPA Mediation Support cell, were already using such tools, while others remained cautious about how technology might interfere with delicate human processes. He mapped the work onto three phases—conflict prevention, dialogue and mediation, and post-conflict peacebuilding—and pointed to concrete openings in each: the ability to crunch multidimensional data to strengthen early-warning and monitoring systems; deliberative technologies that widen participation and buy-in so agreements are not merely elite bargains; help with language and interpretation across divides; and the surfacing of common positions and shared needs. Implementation after a deal, he stressed, is the harder task, and scenario planning tools could support the difficult decisions it requires. Yet he cautioned that speed, so often celebrated as AI’s great virtue, can be the wrong thing in a process that moves at a human pace, and that trust must be built through human contact rather than through machines.
On how such tools should be developed and scaled, Weinberg resisted the idea of a single global “AI diplomat” designed by technologists in one room. He saw value in a deep-knowledge resource trained on the accumulated wisdom of those who have worked through conflict processes, but insisted this could never solve conflict on its own. Any real work, he argued, must be locally led and context led, innovative and creative, and ultimately grounded in people talking to one another.
“What if we could curate something that was the ancestral intelligence—not the artificial, but the ancestral intelligence of our community and wisdom?”
John, Lord Alderdice
Lord Alderdice said he had accepted the invitation to the Turing roundtable more to listen than to instruct, on the conviction that the sooner one engages with a fast-moving technology, the better one can keep pace with it—much as he had opened an online banking account early, years ago, simply to stay abreast of developments. His deeper reason was professional: having run Oxford’s Changing Character of War Centre, he had watched AI reshape conflict itself, including the identification and targeting of people. If the technology of war was advancing, he argued, the technology of peacebuilding had to keep pace, both to seize new capacities—such as drawing together and mapping data—and to understand how other parties to a conflict were using it.
He was candid about the reservations. Some are shared with every field: in one project of his own, the crucial concern was not the AI but where the data would be processed, since material handled in the United States raised questions of confidentiality and trust. Others are graver. He described the work of a colleague, Sundeep Waslikar, who has been convening senior figures from several countries in Geneva to draw up guardrails against the use of AI in nuclear-weapons decisions, where hypersonic delivery leaves vanishingly little time for human judgment. These examples, he noted, cut both ways: the very dangers can prompt thoughtful people to build protections that bring adversaries together.
Alderdice returned repeatedly to two limits. The first is speed: AI can process vast data in moments, but human beings cannot be sped up in the same way, and conflict is bound up with emotion and disturbed relationships that no machine resolves. The second is ethics and ownership. He recounted American colleagues with the technical capacity to build such systems who nonetheless sought independence from what they saw as the questionable ethical standards of large corporations. Ethical disagreement, he argued, is not simply a matter of right and wrong but of genuine divergence between cultures—an LLM trained in Silicon Valley differs from one trained in Beijing, and Western individualism differs from the communal emphasis of much of the East. He also flagged mounting evidence that heavy technology use shapes how young people develop emotionally and cognitively, citing the restrictions being placed on children’s phone use in Australia.
“There is not a global agreed ethical position on anything. Isaiah Berlin was very clear: there is no agreement on the good.”
Pieter François
François, who both moderated and joined the exchange as a scholar of cultural evolution, framed the session by narrowing its focus. AI now touches every aspect of modern life, he observed, and conflict is everywhere; one could examine how AI is reshaping the conduct of war or the environmental strains that set the conditions for conflict. He had instead brought together an experienced negotiator and an experienced facilitator to concentrate on a narrower question: how AI is affecting the peace negotiation process itself.
He was especially drawn to AI’s potential to foreground common ground. Peace processes, he noted, fixate on differences—parties are drawn into what divides them and into how those divisions might be bridged—yet what adversaries already share, even before they reach the table, is easily lost from view. Since human beings are naturally pushed toward difference, he saw real promise in tools that could surface shared positions as a counterweight.
François also pressed on the mismatch of pace between the technology and our understanding of it, joking that anyone rewatching the webinar in a year would find it dated. On surveillance, he agreed with the speakers that the topic was somewhat adjacent: its data captures the imagination, but it is only one of many, far larger datasets that could inform an AI-assisted negotiation, and no such process need depend on it.
“The technical capacity is developing so quickly, but the ethical implications, the biases, the cautions are much more fundamental—and they develop far more slowly.”
Question and Answer Session
How do we reconcile AI’s potential for peacebuilding with its environmental costs?
Weinberg acknowledged the issue as a huge and genuine one. He hoped that the scale of investment might drive innovation toward less damaging systems, and that the tools—which have shown real promise, as in medicine—could be pursued responsibly, but he insisted the risks were real and must be openly discussed. Alderdice added that events like the webinar matter precisely because they alert people to such challenges: the aim is not to frighten anyone into inaction, but to prompt those involved to ask what can be done rather than dismiss the problem because there is money to be made or flee it because it feels dangerous.
What role does AI play in surveillance, and how useful is surveillance data for peace negotiation?
Weinberg turned the question toward opportunity, noting that AI could support transparent verification and the monitoring of ceasefire lines and agreements, while framing surveillance itself as a security discussion outside his expertise. Alderdice addressed the ethical hazard directly. Automating intelligence-gathering—connecting thousands of cameras to AI that analyzes the faces of ordinary people not suspected of anything—invests enormous power in whatever authorities hold it, and demands guardrails. Just as societies once regulated the reach of radio, television, and cinema, he argued, the pairing of AI with surveillance is among the most important and constant dangers to watch.
What must change in humanity for AI to help us build peace?
Alderdice offered a practical illustration. A colleague at Oxford has been gathering data on the negotiation of constitutions—beginning with the American Constitution and extending to conflicts and their constitutional resolutions worldwide—and making it open source. AI can compile and make such material available, along with techniques of negotiation of the kind Weinberg described. This does not solve the underlying problems, he said, but it supplies a genuinely helpful set of resources, made possible by AI’s extraordinary capacity to analyze and share knowledge.
Concluding Remarks
Asked whether AI is, on balance, good or bad for peace, Weinberg suggested reversing the framing altogether: the conversation should be about peace and AI, or conflict and AI, because people—not the technology—are the main actors. Given that AI will be used to prosecute war, he argued, peacebuilders have a responsibility to understand these tools and to work creatively with the technology sector; the dullest use of AI would be simply to make existing jobs easier, while the most exciting is to ask whether it can help address the hardest and most consequential challenges.
Alderdice, declining to reduce the matter to a single verdict, judged AI a serious short-term challenge for both conflict and the environment, and likened it to the invention of printing five centuries ago—a technology that at first fueled the wars of religion, yet one no one would now wish undone. His longer view was hopeful.
“But over a period of time, I have some optimism that humanity will find a way to enable progress to come out of this, and to enable us to deal better with some of the challenges that we have.”
François thanked both speakers, framing the exchange not as a solution but as an ongoing conversation that will continue at the G20 Interfaith Forum in the years ahead.
———
JoAnne Wadsworth is a Communications Consultant for the G20 Interfaith Forum Association and Editor of the Viewpoints Blog, IF20’s platform for commentary at the intersection of faith and global policy. She has covered the work of the Forum across multiple G20 cycles, producing summaries, articles, and editorial content that bring the voices of interfaith leaders and scholars to a broader audience. Her writing engages with topics spanning environmental governance, religious freedom, anti-racism, and the role of faith in multilateral policy processes. She is based in the United States.