A New Push for Human Oversight in Military AI Weapons
A bipartisan group of lawmakers unveiled new legislation on Friday aimed at ensuring that human decision-making remains a central aspect in the use of AI-powered military systems designed for lethal operations.
Representatives Tom Barrett (R-MI), Don Beyer (D-VA), and Sarah Jacobs (D-CA) have put forward the Human Authority Over Autonomous Weapons Act. This proposed legislation seeks to mandate that human oversight be legally required whenever the U.S. military employs autonomous, AI-driven weapons systems.
The bill stipulates that the Department of Defense must guarantee that any intentional lethal actions taken by these AI weapons have human approval or follow established human-involved protocols. This requirement will be necessary throughout military operations where these autonomous systems may potentially be involved in lethal actions.
Additionally, military commanders will be obliged to validate AI-generated targets through non-AI methods for a period of five years post-enactment. This verification is intended to act as a safeguard against possible mistakes and unintended results from AI targeting systems. It’s worth noting that missile defense systems are excluded from these requirements.
The initiative comes as lawmakers from both parties aim to create a legal framework for the military’s engagement with AI, especially with the growing integration of autonomous weapons in contemporary combat scenarios. While the Department of Defense already enforces policies that require a degree of human judgment regarding the use of force, this new legislation would turn these oversight provisions into federally binding laws.
The introduction of this bill follows significant incidents in military artificial intelligence. For instance, the Pentagon utilized AI in a raid that captured Venezuelan President Nicolás Maduro, which led to tensions with Anthropic, a company expressing concerns over the boundaries around autonomous weaponry and mass surveillance, resulting in a legal conflict. This situation underscored worries about the implications of AI in military applications and the ethical dilemmas that arise.
In April, a coalition of religious leaders urged Congress to mandate substantial human control over AI-equipped weapon systems, reflecting wider societal apprehensions about autonomous military technology.
Supporters of the bill highlighted both ethical and pragmatic reasons for its introduction. Representative Beyer noted, “No machine should have the authority to decide to kill humans independently,” while Barrett emphasized the need for proactive measures to ensure the ethical and transparent use of these technologies.
AI-driven autonomous weapons have gained attention, notably in a book titled Code Red: Left, Right, China, and the Race to Control AI published earlier this year. The author discusses how these weapons could be likened to the nuclear bombs of this century.
The author notes that the availability of lethal AI weaponry is democratizing, meaning that technologies once only accessible to superpowers will soon be available to a wider array of actors, both state and non-state.
Indeed, the stakes involved are immense; as great powers may resort to more destructive weaponry. For example, Israel’s use of AI during its conflict in Gaza illustrated how machine learning can analyze data and identify targets more rapidly than human operators.
Connecting these sophisticated data systems to autonomous weapons brings humanity closer to a perilous future known as “hyperwar,” a term indicating an all-out conflict among technologically advanced nations. Proponents argue that these autonomous systems’ speed could overwhelm adversaries in fast-paced battle scenarios.
