人工智慧對國際和平的機遇與挑戰

Opportunities and Challenges of Artificial Intelligence for International Peace

本文旨在探討人工智慧對國際和平建構的影響，分析其作為技術工具的潛力與風險。隨著和平研究從傳統的權力平衡論演進至涵蓋「積極和平」與「混合和平」的多元觀點，人工智慧技術的介入正改變國際社會應對衝突的模式。研究指出，人工智慧透過巨量資料分析與模式辨識，能將衝突預警從被動回應轉為主動預防；利用自然語言處理工具，輔助和平談判並擴大邊緣社群的參與，促進包容性對話；透過衛星影像與地理資訊系統的整合，提升維和行動監測與人道援助資源配置的精準度與效能。另方面，演算法偏差可能再現結構性暴力，導致弱勢群體落入演算法的識別盲區；決策過程的黑箱化造成課責困境，削弱制度正當性；資料與算力的集中引發資料殖民與權力不對稱，弱化在地主體性；資訊污染與監控濫用侵蝕社會互信。本文發現，人工智慧賦能和平的關鍵不在於技術本身，而是如何將其嵌入可課責、可協調且可檢驗的制度框架，因此本文建議架構以人機協作為核心的歸責機制，校正演算法的文化與地域偏差，並透過多邊協調緩解全球治理碎片化，使人工智慧成為建構國際和平的有效工具。

This study explores the impact of Artificial Intelligence (AI) on international peacebuilding, analyzing its potential as a technological tool alongside its associated risks. As peace studies evolve from traditional balance-of-power theories to diverse perspectives encompassing ''positive peace'' and ''hybrid peace,'' the intervention of AI technology is transforming how the international community addresses conflict. The research indicates that AI, through big data analysis and pattern recognition, can shift conflict early warning from passive response to proactive prevention. Using Natural Language Processing (NLP) tools, AI assists in peace negotiations and expands the participation of marginalized communities, fostering inclusive dialogue. Furthermore, the integration of satellite imagery and Geographic Information Systems (GIS) enhances the precision and effectiveness of monitoring peacekeeping operations and allocating humanitarian aid resources. On the other hand, algorithmic bias may reproduce structural violence, leading to the systemic exclusion of vulnerable groups. The ''black-boxing'' of decision-making processes creates accountability dilemmas, weakening institutional legitimacy. The concentration of data and computing power triggers data colonialism and power asymmetries, undermining local agency. Moreover, information pollution and the abuse of surveillance erode social mutual trust. This paper finds that the key to AI-empowered peace lies not in the technology itself, but in how it is embedded within an accountable, coordinated, and verifiable institutional framework. Therefore, this study suggests establishing an accountability mechanism centered on human-AI collaboration, correcting the cultural and local biases of algorithms, and mitigating global governance fragmentation through multilateral coordination, making AI an effective tool for building international peace.