Robot diplomacy: How AI could usher in a new era of world peace
The world is facing the most significant surge in violent conflicts since World War II. According to a United Nations briefing, a quarter of the global population is affected by conflicts today, the likes of which are more complex than ever.
This new era of rising conflict rates demands an honest reexamination of how we do things. It is imperative for practitioners in the fields of peace and defense to acknowledge that conventional approaches are failing one-fourth of humanity. Disruption is urgently needed, and artificial intelligence (AI)-powered peace should be our north star.
AI-powered peace has the potential to revolutionize international peace and security. It will allow practitioners and policymakers to eliminate the time and resource intensity behind data collection, analysis and option generation, and free up those resources to focus on the difficult things like dialogue, negotiation, trust-building and strategic decisionmaking.
Machine learning and generative AI can learn from data, identify patterns, predict outcomes and generate new knowledge and solutions.
Researchers have developed AIs that can detect cancer sooner than physicians currently can. This advance makes early-stage treatment more likely and will ultimately reduce cancer deaths. Likewise, when it comes to the vast realm of challenges and unknowns faced by national governments and multilateral organizations, AI will not only organize the information cacophony in seconds, but also illuminate a symphony of ideas and strategies that have yet to be discovered. With the right decisions made now on AI, we could be at the doorstep of world peace within our time.
AI-powered peace holds great promise and has already shown meaningful effectiveness in various applications. But the technology remains underdeveloped and under-used. In Libya, for example, AI-assisted large-scale digital dialogues advanced the peace process by facilitating a conversation between 1,000 Libyans at key milestones of its implementation. This helped create a consensus that helped establish an interim government in just four months.
In contrast, traditional dialogues often take years to complete, especially in the most violent countries with hard-to-reach populations.
The digital dialogues applied a machine learning model with natural language processing capabilities that could translate and group emerging interests in real-time. Today, Libya has moved up to 137 out of 163 countries in the Global Peace Index ranking, marking a significant improvement from its 156th position in 2020. Despite challenging circumstances, Libya has also achieved the largest overall improvement in peacefulness worldwide for the second consecutive year.
There is an untapped gold mine to be found in AI applications for peace, as they are possible across conflict cycles and timelines. Opportunities exist in the development of conflict analyses, mediation and negotiation support, and peacekeeping, ceasefire monitoring and reporting, and more. Natural language processing can be used across peace processes to facilitate thousands of citizen responses in their local dialects to open-ended questions on their security sector, governance, constitution, transitional justice arrangements and more.
The integration of this technology would revolutionize how citizens engage in the building of their nations. Peace processes could be faster, more inclusive, more responsive, less violent, and overcome cross-language communication barriers. Further research and development in machine learning and natural language processing are essential to work with less-spoken languages and address low literacy rates, restricted technology and internet access, and overall accessibility.
Sentiments analysis is another AI application that could be invaluable to advancing peace. This facet of natural language processing has already been pioneered in the private sector, where companies employ it to discern the emotional tone surrounding a subject — positive, negative, or neutral.
This capability is exceptionally valuable, particularly considering the U.S. government’s strong emphasis on understanding the viewpoints of populations in conflict zones. Sentiments analysis could be applied anywhere that natural language processing is available. It brings the added benefit of understanding reactions to proposed interventions or current events to inform decision-making.
Such applications of AI machine learning can shed light on a nation's most ambitious and elusive goals faster than ever before while saving lives. For example, while desired forms of governance often emerge at a macro level in peace negotiations, fleshing out the exact details to reach a final settlement and then implement it can be a protracted and expensive process. In the meantime, violence can prevail as the only means of determining who holds power and authority. Machine learning applied throughout a peace process could expedite consensus on a multitude of governance arrangements within weeks, ultimately saving lives and averting further deterioration in health, food access, and economies.
Large Language Models (LLMs), made famous by Open AI’s Chat GPT, also have vast applications to the field of peace and security. Consider conflict and strategic analysis, a critical and foundational tool in the field that informs all decisions and recommendations. For this product to be strong, it must have a breadth of sources and data and accurately reflect up-to-date citizen and stakeholder interests.
Sometimes, decisionmakers are forced to respond rapidly to emergencies despite remaining gaps in analysis. Rather than rely on lengthy conventional analytical methods and accept their shortcomings, LLMs can synthesize existing text — research, reports, social media, and new queries — and complete such analyses in minutes. The use of LLMs can thus position analysis and decision-making at the forefront of innovation.
AI can also apply machine vision and image recognition to transform peacekeeping and ceasefire monitoring and reporting. These technologies are already mastered within the defense and private sectors through image recognition, autonomous drones, and self-driving vehicles for example. UN peacekeeping missions and ceasefire monitoring mechanisms could surveil and report incidents without compromising human lives. Moreover, autonomous monitors and reporting could alleviate the funding shortfalls that these institutions perpetually face.
It is time is to develop and leverage formidable AI capabilities for peace. But currently, there is no funding for it. The U.S. Department of Defense has allocated $1.8 billion for AI in its 2024 budget. What if a portion of that could be reoriented toward AI-powered peace? It is essential for the U.S. to continue to open investment channels and research and development for this purpose across government, non-profit, and private sectors.
With the most advanced AI economy in the world, the U.S. has a unique opportunity to lead. The chance to revolutionize peacemaking and achieve a world without violent conflicts in our lifetime is very real. Embracing AI for peace is not merely a choice — it is a necessity, and the time to do it is now.
Ola Mohajer was a refugee and is a survivor of war. She is a senior program officer at the United States Institute of Peace and is exploring building AI for peace at missiontranscend.ai. Her views do not necessarily reflect those of either organization.