U.S. News

Warnings of tech apocalypse as Pentagon amps up AI-weapons program

By Jake Beardslee · October 1, 2023

In brief…

The Pentagon's plan to rapidly deploy autonomous weapons through its new Replicator initiative has drawn criticism from human rights groups.
Critics warn of an uncontrolled AI arms race that could place humanity at risk.
Verifying compliance with any treaty or oversight program will be challenging as advanced AI can evade monitoring.
Some experts liken the advent of automated weapons systems to the dawn of the nuclear weapons era.

The Pentagon's rapid deployment of autonomous AI weapons systems raises urgent ethical concerns and fears of an out-of-control arms race. Air Force Staff Sgt. Brittany A. Chase, DOD/Wikimedia

The potential of “killer robots” powered by artificial intelligence (AI) has critics calling for far stronger guidelines with fears that a tech-fueled arms race could one day overrun humanity.

The Pentagon’s new Replicator initiative aims to rapidly deploy thousands of fully autonomous weapons systems over the next two years. With no international treaties in place, human rights groups are calling for urgent action to avoid potential catastrophe as algorithms increasingly displace human decision-making.

While self-operating drones and aircraft have existed for decades, recent advances in AI allow for swarms of interconnected systems that can function with little human oversight.

While the Pentagon claims it will retain “appropriate” control, vague ethical guidelines and broad waivers have raised red flags. Experts warn unchecked proliferation of these systems could lead to unintended warfare triggered by automated, non-human computer codes.

The UN has debated such regulations since 2013, but has not yet reached a consensus. There may soon be progress on that front as the General Assembly has recently begun to increase its focus on the issue of autonomous weapons systems.

While the U.S. has not committed to any treaty, verifying compliance will be difficult, since advanced AI can evade monitoring. Activists have demanded a ban on enabling robots to kill an adversary without a human overseer, citing the AI’s inability to value life.

With only one known prior combat use-case, questions remain about the real-world impacts of mass deployment of these potentially apocalyptic systems. In response, the Pentagon has formed an ethics working group and oversight office

Meanwhile, worries persist as some experts liken the advent of automated weapons systems to the dawn of the nuclear weapons era.