The United Nations has commenced its inaugural Global Dialogue on AI Governance in Geneva to address the rapid proliferation of artificial intelligence and the necessity for international regulatory synchronization.

聯合國在日內瓦啟動了首次人工智慧治理全球對話，旨在應對人工智慧的快速普及以及國際監管同步化的必要性。

The current discourse is predicated on the observation that AI development is outpacing the adaptive capacity of sovereign governments and scientific understanding. Secretary-General António Guterres has posited that the transition of civilian-grade AI hardware to military applications has facilitated the emergence of lethal autonomous weapon systems, which he asserts must be prohibited under international law. This systemic risk is compounded by the potential for AI to destabilize global security and societal structures, a concern first articulated by the Secretary-General in 2017.

目前的討論是基於一個觀察，即人工智慧的發展速度已超過主權政府的適應能力與科學理解。秘書長古特雷斯指出，民用級 AI 硬體轉向軍事應用，促成了致命自主武器系統的出現，他主張根據國際法必須禁止此類系統。這種系統性風險因 AI 可能導致全球安全與社會結構不穩定而進一步增加，秘書長早在 2017 年就首次提出了這一擔憂。

Central to the proposed governance model is the protection of vulnerable populations, specifically children. The administration has advocated for an 'AI Child Safety Pledge,' which would mandate independent oversight and rigorous safety testing prior to deployment. This is necessitated by data indicating that a vast majority of deepfake content is sexualized and disproportionately targets women and girls. Furthermore, the UN emphasizes the preservation of human agency in high-stakes sectors such as jurisprudence, healthcare, and policing, asserting that while AI may provide data, final decision-making authority must remain with human actors.

擬議治理模式的核心是保護弱勢群體，特別是兒童。聯合國倡導一項「AI 兒童安全承諾」，要求在部署前必須進行獨立監督與嚴格的安全測試。這是由於數據顯示，絕大多數的深偽內容均為色情化，且不成比例地針對女性與女童。此外，聯合國強調在法理、醫療和警務等高風險領域需保留人類的主體能動性，主張雖然 AI 可提供數據，但最終決定權必須留在人類手中。

Institutional efforts are also directed toward mitigating the 'AI divide.' The Secretary-General highlighted a significant disparity between private investment, estimated at $500 trillion, and the negligible public funding available to developing nations. To facilitate a rapprochement between technologically advanced and developing states, the UN has proposed a Global Fund for AI and a Global Network for Exchange and Cooperation on AI Capacity Building. Additionally, environmental sustainability is identified as a critical metric, with a requirement for all AI data centers to utilize renewable energy by 2030 to mitigate substantial carbon and water footprints.

機構努力亦指向減輕「AI 鴻溝」。秘書長強調，私人投資（估計達 500 兆美元）與開發中國家獲得的微小公共資金之間存在顯著差距。為了促進技術先進國家與開發中國家之間的趨同，聯合國提議成立「全球 AI 基金」以及「全球 AI 能力建設交流與合作網絡」。此外，環境永續性被視為關鍵指標，要求所有 AI 數據中心在 2030 年前使用再生能源，以減輕龐大的碳足跡與水足跡。

Stakeholder positioning varies regarding the timeline of implementation. While the UN seeks a comprehensive global agreement, General Assembly President Annalena Baerbock has argued that the urgency of protecting children and preventing the use of AI for genocide warrants the immediate adoption of safeguards by coalitions of states, rather than awaiting a full consensus among all 193 member states.

各持份者對於實施時間表的立場不一。雖然聯合國追求全面的全球協議，但大會主席貝爾博克認為，保護兒童與防止 AI 被用於種族滅絕的迫切性，使得由國家聯盟立即採取保障措施更為重要，而非等待所有 193 個會員國達成完全共識。

The United Nations continues to coordinate international stakeholders to establish safety standards and equitable access, with a subsequent dialogue scheduled for May 2027 in New York.

聯合國將繼續協調國際持份者以建立安全標準與公平獲取權，後續對話預計於 2027 年 5 月在紐約舉行。