The warning from the Pentagon comes as both countries look to bring through new artificial intelligence-enabled surveillance technologies to keep tabs on their populations China and Russia are both developing autonomous drones with the potential to kill without any human involvement, US military chiefs have warned.
Pentagon leaders in charge of America’s military applications of artificial intelligence claim both countries are ignoring the ethical limits or considerations regarding the use of such drones
Artificial intelligence is increasingly being used in both civilian and military operations. The comments came from Nand Mulchandani, Acting Director of the Joint Artificial Intelligence Center, according to a Pentagon transcript.
“We know that China and Russia are developing and exporting AI-enabled surveillance technologies and autonomous combat systems without providing any evidence of adequate technical or ethical safeguards and policies,” he told reporters.
Mulchandani highlighted China’s use of AI for domestic policing issues such as censorship and facial recognition, but also warned of many military applications which are causing alarm.
This month in China, for instance, officials deployed drones to fight locust swarms in ravaged farmland.
He said certain ethical constraints may not exist for Russian or Chinese weapons developers, who could quickly leverage different kinds of warfighting robots.
For example, armed drones could advance on enemy soldiers without needing a human to operate them, and use built-in sensors to avoid obstacles, find targets and even fire weapons autonomously.
National Interest reports the Pentagon has continued to use its existing so-called “man-in-the-loop” doctrine. This means that all decisions regarding the use of lethal force by unmanned vehicles, drones and other systems must be made by a human being.
Rapid advances in autonomy and AI make it possible for such systems to detect, track, and destroy targets without any human involvement at all.
As well as the ethical issues regarding using autonomous systems to kill, there are also technical reasons why the Pentagon insists on human involvement.
An AI system is only as effective as its database, meaning certain algorithms may not yet have sufficient reliability to accurately identify everything that is detected.
In addition, enemies can work to “spoof” or “confuse” certain sensing algorithms.
US military scientists and AI developers often explain there are a few defensive uses of AI, which could be used to fire interceptors and detect and destroy incoming rockets, missiles, artillery or drones.
These applications, while already under development in areas such as ballistic missile interceptors and drone swarm defence, are not the same as offensive uses of lethal force against enemies.