The article argues that the reported Iranian capture of an Anduril Dive-LD unmanned underwater vehicle highlights major gaps in how the U.S. thinks about autonomy in military systems. Although the Dive-LD is described as an attritable autonomous underwater vehicle designed for intelligence, surveillance, mine warfare, and other risky missions, its loss matters because a complete captured system could reveal operational details, weaknesses, and potentially sensitive technologies to adversaries such as Iran, China, or Russia.
The author says the incident exposes limits in Department of Defense Directive 3000.09, which focuses on safety, cyber survivability, anti-tamper measures, and minimizing unintended engagements. But the directive does not fully address real-world problems such as malfunction, capture, contested communications, jamming, spoofing, and the possibility that an autonomous system may need to make decisions when human oversight is unavailable.
A central theme is “bounded autonomy”: low-risk missions could be fully autonomous, while high-risk missions should require stronger human control. The article stresses that autonomous systems are not people and cannot be held morally accountable, so commanders remain responsible for their actions. It also raises a difficult question about whether autonomous vehicles should be programmed to scuttle or zeroize themselves if capture seems imminent, since such safeguards could also be exploited by enemies.
Overall, the piece argues that autonomy in warfare requires clearer rules about delegation, failure, and control before future conflicts make these problems unavoidable.
Article timeline
Build this article’s timeline
Extract dated events, key actors, and open questions using the same timeline analysis available in Signal.
Comments
Login to view and post comments.