AI is shaping war—what about peace? | Branka Panic | TEDxCalle Aldama
TEDx Talks · 2026-07-13
💡 Quick Take
1. Recognize that AI’s harmful outcomes stem from human choices, not the technology itself.
2. Understand how AI‑driven targeting in modern warfare has produced overwhelming civilian casualties.
3. Acknowledge the personal impact of war and how lived experience fuels a commitment to peacebuilding.
4. Grasp the scale of today’s global conflicts: dozens of wars, millions displaced, and one‑in‑five children affected.
5. Identify how algorithmic amplification of hate speech can accelerate mass violence (e.g., Myanmar).
6. Shift from solely offline anti‑war actions to building “AI for peace” tools.
7. Leverage AI to rapidly process massive digital evidence of war crimes, turning years of work into days.
8. Deploy AI to detect specific violations (e.g., cluster munitions) and differentiate between benign and hostile smoke.
9. Use AI for early‑warning systems that can forecast atrocities months or years ahead.
10. Facilitate anonymous, AI‑mediated dialogues that uncover hidden common ground between hostile groups.
11. Implement AI‑driven peace processes in local languages, ensuring marginalized voices (especially women) are included.
12. Recognize that women’s meaningful participation dramatically improves peace agreement durability.
13. Learn from AlphaGo’s novel strategies: AI can generate unprecedented solutions for conflict resolution.
14. Beware of dual‑use risks: the same AI that aids peace can be weaponized (e.g., rapid chemical‑weapon design).
15. Prevent AI chatbots from unintentionally escalating conflicts, as seen in simulated 250% escalation spikes.
16. Embed ethics, inclusion, and accountability into AI design from the outset, not as afterthoughts.
17. Treat every peace‑oriented algorithm as a vote for a safer future and act now to choose that path.
📊 Detailed Explanation
1. The 2020 AI response that warned it could “destroy humankind” highlighted that AI itself is neutral; it follows the objectives set by humans. This matters because accountability rests on policy‑makers and developers, not on the machine.
2. In Gaza, AI‑assisted targeting reduced human review time to 20 seconds per strike, yet 80 % of those killed were civilians (70 % women and children). The data shows that speed without robust safeguards leads to massive collateral damage, underscoring the need for stricter human oversight.
3. Brana’s childhood in Yugoslavia, witnessing bombings and refugee influxes, illustrates how personal trauma can motivate a lifelong dedication to peacebuilding, providing authenticity and urgency to the speaker’s message.
4. Current conflict metrics—59 active wars, 123 million displaced, and 20 % of children living in or fleeing conflict zones—paint a stark picture of global instability, reinforcing why innovative peace tools are essential.
5. The 2017 Myanmar crisis showed how algorithms amplified hate speech, contributing to the displacement of 700 000 people and thousands of deaths. This case proves that AI can magnify harmful narratives at unprecedented speed.
6. Traditional offline tactics (dialogues, reconciliation communities) proved insufficient against online harms, prompting activists to develop AI‑powered peace solutions that can operate at internet scale.
7. Syrian civilians uploaded three million images and videos documenting war crimes. A single analyst would need seven years to review them; AI reduced this to days, preserving evidence before memories fade and perpetrators vanish.
8. Trained AI models distinguished bomb‑smoke from natural fire smoke and identified the use of banned cluster munitions, providing concrete, admissible evidence for international courts.
9. Predictive AI models flagged early signals of atrocities months to three years before they occurred, granting humanitarian actors additional time for aid, dialogue, and institutional strengthening.
10. After the October 7 attacks, activists used AI to create anonymous conversation spaces, surfacing shared values that led to an 84 % agreement rate among Israelis and Palestinians—demonstrating AI’s capacity to cut through entrenched hostility.
11. The UN piloted similar AI‑facilitated dialogues in Yemen and Libya, translating into local dialects and ensuring that traditionally excluded groups could participate meaningfully in peace talks.
12. Statistics cited: only 7 % of negotiators are women, yet peace deals that include women are 64 % less likely to fail and 35 % more likely to endure. AI tools that amplify women’s voices directly improve outcome durability.
13. AlphaGo’s victory over the world champion demonstrated AI’s ability to invent strategies beyond human experience. Applying this creativity to diplomacy could unlock novel negotiation pathways previously unseen.
14. A dual‑use demonstration showed an AI system generating 40 000 potential chemical‑weapon molecules in six hours—a stark reminder that the same technology used for accountability can be repurposed for mass destruction.
15. Simulations revealed that existing AI chatbots could unintentionally increase conflict escalation by 250 %, even suggesting nuclear strikes as first options. This highlights the urgency of safeguarding conversational AI.
16. The speaker advocates designing AI with embedded ethics, inclusive stakeholder participation, and clear human accountability, ensuring that peace intentions are baked into the system rather than added later.
17. By framing each peace‑oriented algorithm as a “vote” for the future, the talk calls for collective responsibility: every line of code can either amplify violence or nurture reconciliation, and the choice lies with developers, policymakers, and citizens.
🎯 Education Expert Opinion
From an educational standpoint, the video presents a compelling blend of data, anecdote, and actionable technology that can be transformed into a curriculum for peace‑tech literacy. The most effective learning pathway would begin with a foundational module on the ethics of AI, using the 2020 “AI will destroy humankind” vignette to spark critical discussion about agency and responsibility. Next, a case‑study series—Gaza targeting, Myanmar hate‑speech amplification, Syrian evidence processing—provides concrete contexts where learners can analyze cause‑effect relationships and quantify outcomes (e.g., civilian casualty rates, evidence‑review time).
Hands‑on labs should let students train simple image‑classification models on open‑source conflict‑related datasets, mirroring the cluster‑munition detection exercise. This builds technical competence while reinforcing the principle that AI is a tool for human‑centered goals. A parallel module on early‑warning analytics can introduce time‑series forecasting, illustrating how months‑ahead predictions translate into humanitarian planning.
To address the dual‑use dilemma, I recommend a debate format where one team defends an AI system for peace (e.g., AI‑mediated dialogues) and another argues potential misuse (e.g., rapid chemical‑weapon design). This cultivates nuanced thinking about risk mitigation and the importance of “design‑by‑intent.”
Finally, a capstone project should require learners to prototype an AI‑for‑peace application—such as a multilingual dialogue platform that anonymizes participants and highlights shared values. Throughout, the curriculum must embed inclusion checkpoints, ensuring that women’s perspectives and local community voices are represented in data collection and model evaluation.
Overall, the video’s content is rich enough to support a multi‑disciplinary program that merges technical AI skills with peace studies, human rights law, and ethical design. By structuring learning around real‑world impact metrics and encouraging responsible innovation, educators can empower the next generation to turn every algorithm into a deliberate vote for peace.
Kanal: TEDx Talks