Image Source: University of Pennsylvania
Our ways of waging war have evolved; it is time our ways of waging peace do as well.
It’s been over a year since Russia started its war against Ukraine, and it’s not likely to end anytime soon. Despite the odds, Ukraine has managed to hold its ground, thanks in no small part to cutting-edge technologies. Aerial drones and open-source AI systems have given Ukraine an edge on the battlefield.
But although we’ve made advances in how we fight wars, we haven’t managed similar strides in peacemaking. Wars don’t end—at least, not peacefully—with victories on the battlefield. They end with negotiations, and when it comes to diplomacy, we’re stuck in the 19th century. Ukraine has shown us how crucial innovation is in the face of aggression, but when the time comes for negotiations, we will need all the technological help we can get to secure a lasting resolution.
Enter ChatGPT. The AI language model trained by OpenAI has been making waves in the tech world since its soft release in November. ChatGPT cannot replace human diplomats, but by working in tandem with AI, we can explore new and innovative ways to bring about a peaceful resolution to conflicts like the one in Ukraine.
ChatGPT stands apart from other AI tools in a way that is valuable for conflict prevention and resolution. Its specific behavior guidelines are designed to avoid inflammatory or divisive content. As a result, ChatGPT is more likely to produce neutral language fit for negotiations or conflict de-escalation.
We’re already seeing how AI tools can make a difference in diplomacy. Google’s language-translating glasses can interpret text in real-time, bridging communication gaps and fostering cultural exchange. ChatGPT, too, has the potential to break down cultural barriers when translators aren’t readily available. We asked ChatGPT how to say “What are you doing?” to a formal audience in Telugu, a Dravidian language native to southern India, and here’s what it gave us:
Looking side-by-side, Google simply translates the text and defaults to informal language, but ChatGPT has the potential to do much more. We can ask ChatGPT to break language down into comprehensible pieces and communicate cultural nuances specific to a particular language.
Google Translate and similar tools have provided basic translation services for some time, but they may fall short in capturing the subtle nuances and cultural complexities that are integral to effective diplomacy. ChatGPT, on the other hand, is designed to break down language barriers and communicate cultural subtleties. In doing so, it could surpass other tools—perhaps even human translators—in several key ways.
Firstly, the speed and easy availability of ChatGPT make it a valuable resource for diplomats working across different time zones or in time-sensitive situations. Real-time translations allow for more efficient communication, which is crucial in crisis scenarios.
Secondly, scalability is a major advantage of ChatGPT. Handling large volumes of text and multiple languages is not an issue for the AI, whereas human translators may struggle to keep up or maintain consistent quality throughout. Finally, the AI is designed to continually learn and improve its understanding of languages and cultures. It is true that ChatGPT’s current capabilities may not produce a perfect answer or translation every time. Yet the AI’s growing database of cultural knowledge and learning ability to adapt to new contexts and scenarios could eventually surpass an individual human translator’s cultural exposure.
Now, let us take it a step further. Remesh is an AI start-up that’s been working with the United Nations to broker peace deals in Yemen and Libya. The Remesh tool gathers data from people in conflict zones, helping the U.N. understand their concerns in real time. GPT-4—ChatGPT’s latest iteration—could go above and beyond this by processing images in addition to text. GPT-4 could analyze warzone photographs from Yemen and Libya, enhancing policymakers’ understanding of the humanitarian situation on the ground.
AI platforms can also provide data-informed advice during negotiations, leading to better decisions and quicker agreements. IBM’s Cognitive Trade Advisor is one tool that has already helped negotiators by answering complex trade treaty questions. ChatGPT has an even broader scope, capable of exploring issues beyond trade. One of the key advantages of ChatGPT lies in its capacity to analyze vast amounts of data rapidly, enabling negotiators to make informed decisions based on historical records, previous negotiations, and media reports. This data-driven approach can inform strategies and help negotiators address complex legal issues with greater confidence and accuracy.
One of the most significant ways ChatGPT can assist negotiators and diplomats is by analyzing lengthy legal documents and distilling them into concise, easily digestible summaries. By translating these intricate documents into more straightforward language, ChatGPT helps ensure that even those without extensive legal backgrounds can grasp the core concepts and implications of treaty law, empowering them to engage in more informed discussions.
For example, when asked to summarize and capture the key legal elements of the European Convention on Human Rights, ChatGPT responded:
As the above examples demonstrate, ChatGPT can make the seemingly impenetrable realm of treaty and international law more accessible and consumable for those tasked with shaping global diplomacy and peacemaking. By embracing the unique advantages offered by ChatGPT, negotiators can make more informed decisions, navigate the intricacies of legal frameworks, and ultimately improve their chances of achieving lasting peace agreements.
ChatGPT can also analyze data to prevent conflicts—specifically by detecting patterns that reveal rising tensions and conflict triggers. This is similar to what the U.N.’s innovation cell is already doing with Stanford University and NASA satellite imagery. ChatGPT can go further by integrating advanced language processing capabilities to monitor politicians’ statements and social media discourse in volatile regions, complementing satellite imagery analysis and providing a more comprehensive understanding of potential conflict triggers.
AI platforms like ChatGPT are already being used for real-time disaster relief. For instance, the non-profit GiveDirectly collaborated with Google Fellows to develop the Delphi tool, which maps the intersection of poverty and disaster-related damage on a census block group level. This enables organizations to swiftly pinpoint those most in need, which is crucial in mitigating civil unrest and conflict.
The Delphi tool was employed to analyze aerial photographs taken by NOAA before and after Hurricane Ian hit Florida. Using Google’s SKAI model, a damage prediction technology co-created with the UN World Food Program, Delphi was able to predict damage for hundreds of thousands of structures, expediting the assessment process and optimizing resource deployment.
While the Delphi tool has already made strides in aiding disaster relief, ChatGPT has the potential to go further by weaving together information from diverse sources such as satellite imagery, social media, news reports, and sensor networks. ChatGPT could foster seamless coordination among relief organizations and bridge language barriers with multilingual support.
OpenAI made ChatGPT available to the public very recently, so its true capacity to transform diplomacy and prevent conflict has not yet been tested. ChatGPT’s current iteration is far from perfect, but AI is built to continuously improve. By no means is ChatGPT a substitute for human-centered statecraft, but robots have been aiding us in waging wars for some time—it’s time to start leveraging technology to help us avoid them.
This article was co-authored for GSSR by Zoe Kern and Satya Adabala.