[Karishma Vaswani] AI can’t broker peace

Singapore’s Foreign Minister Vivian Balakrishnan has created an AI-powered tool called NanoClaw, an open-source technology designed to act as a 'diplomatic second brain.' The system curates transcripts, speeches, and other diplomatic materials—particularly Balakrishnan’s own contributions—into a searchable database. He describes it as indispensable, joking that he no longer dares turn it off, though he insists AI cannot replace human judgment in critical negotiations like ending the Ukraine war or resolving US-Iran tensions. Globally, governments are leveraging AI to handle complex diplomatic tasks. The US State Department uses AI for document translation and information summarization, while the UN deployed a platform called Remesh in Libya’s 2020 peace process to gather and analyze citizen input via basic mobile phones, amplifying underrepresented voices. The World Bank is using AI to predict refugee movements from South Sudan and the Democratic Republic of Congo into Uganda, helping governments prepare infrastructure and services in advance. Despite these advancements, risks remain. Asha Hemrajani, a senior fellow at Singapore’s S. Rajaratnam School of International Studies, warns that AI systems are vulnerable to hacking and manipulation, leading to potential strategic errors. The principle of 'garbage in, garbage out' applies, meaning flawed or biased data can produce unreliable outcomes. In high-stakes diplomacy, accuracy is non-negotiable, making human oversight essential. Balakrishnan’s tool raises concerns about job displacement, as it automates tasks traditionally handled by junior diplomats and policy staff, such as drafting briefs and synthesizing reports. Singapore faces particular exposure, with Bloomberg Economics estimating that 40% of its workforce could be affected by AI—the highest share globally. While adoption speed and job augmentation remain uncertain, public sentiment in Singapore is optimistic, with only 17% believing workers are unlikely to benefit from AI, compared to 46% in Germany. The debate highlights a broader tension: AI can enhance efficiency and data processing but cannot replace nuanced decision-making in diplomacy. Balakrishnan’s approach reflects a pragmatic balance—using technology to support human expertise while ensuring accountability remains firmly in human hands.