Recent reports show that Ukraine has used fully autonomous lethal weapons on a limited basis, while more broadly integrating AI to improve how drones navigate and find targets.

最近的報告顯示，烏克蘭在有限範圍內使用了全自主致命武器，同時更廣泛地將 AI 整合其中，用以提升無人機的導航與尋標能力。

Alexander Kokhanovskyy, CEO of Aero Center, mentioned an experimental operation from two years ago where quadcopter drones were programmed to attack targets on their own in a specific area. Although there was no video of the attack, later checks by human operators found dead Russian soldiers, which suggests the autonomous strike was successful. This event shows a move toward weapons that can choose and attack targets without human help, which matches the US Department of Defense's definition of lethal autonomous weapons.

Aero Center 的執行長 Alexander Kokhanovskyy 提到兩年前的一次實驗操作，當時將四軸無人機設定為在特定區域內自動攻擊目標。雖然該次攻擊沒有影片記錄，但隨後由操作員檢查發現有俄羅斯士兵死亡，顯示這次自主打擊取得了成功。此事件顯示武器正趨向無需人類協助即可選擇並攻擊目標，符合美國國防部對致命自主武器的定義。

Despite this, the Ukrainian military generally prefers semi-autonomous systems. Military leaders emphasize that they follow international humanitarian law, ensuring that humans make the final decision to avoid killing civilians. Furthermore, the use of fully autonomous systems is limited because they might accidentally attack their own troops and are difficult to manage within legal frameworks. However, AI is becoming more common for navigation. Kateryna Bondar, a former government advisor, noted that AI-driven navigation has increased the success rate of long-range drones from 20% to 80% by bypassing Russian signal jamming. Similarly, Russia has used Geran-2 drones that use small computers to change targets automatically.

儘管如此，烏克蘭軍方通常偏好半自主系統。軍方領導者強調他們遵循國際人道法，確保由人類做出最終決定以避免殺害平民。此外，全自主系統的使用受到限制，因為它們可能會誤擊己方部隊，且在法律框架內難以管理。然而，AI 在導航方面正變得更加普遍。前政府顧問 Kateryna Bondar 指出，AI 驅動的導航透過避開俄羅斯的信號干擾，將長程無人機的成功率從 20% 提升至 80%。同樣地，俄羅斯使用了 Geran-2 無人機，利用小型電腦自動更改目標。

From a global perspective, these technologies have created a diplomatic problem. For ten years, the United Nations has tried to agree on a ban, but the US, Russia, Israel, and India have blocked these efforts. Additionally, because open-source software and commercial parts are easy to find, it is very difficult to control the development of these autonomous weapons internationally.

從全球視角來看，這些技術造成了外交問題。聯合國嘗試十年地想要達成禁令協議，但美國、俄羅斯、以色列與印度一直阻撓這些努力。此外，由於開源軟體與商業零件容易取得，要在國際層面控制這些自主武器的研發是非常困難的。

The war in Ukraine is speeding up the shift from systems controlled by humans to AI-driven capabilities for navigation and targeting.

烏克蘭戰爭正加速將系統從人類控制轉向 AI 驅動的導航與尋標能力。