San Francisco supervisors have voted unanimously to reverse a previous decision to authorise police to use robots equipped with lethal weapons.
The San Francisco Police Department (SFPD) said it had no plans to arm the robots with guns but wanted the ability to equip robots with explosive charges "to breach fortified structures containing violent, armed, or dangerous subjects".
The policy instantly faced fierce criticism from civil liberties groups, with some saying arming robots was a step too close to something one would see in a dystopian science-fiction movie.
Dr Catherine Connolly, from the group Stop Killer Robots, told the BBC the move was a "slippery slope" that could distance humans from killing.
Three of the city's supervisors were opposed to the motion from the start and joined dozens of protesters on Monday (5 December) outside City Hall to urge the board to change course.
They chanted and held signs with phrases like 'We all saw that movie… No Killer Robots'.
Supervisor Dean Preston was among them, and on Tuesday he told his colleagues the public had not been given enough time to voice their concerns about such a pressing issue.
"The people of San Francisco have spoken loud and clear: There is no place for killer police robots in our city," he said in a statement after the vote. "We should be working on ways to decrease the use of force by local law enforcement, not giving them new tools to kill people."
Hillary Ronen was another of the supervisors who originally voted against deploying killer robots.
"I'm surprised that we're here in 2022," she said during last week's meeting. "We have seen a history of these leading to tragedy and destruction all over the world."
After Tuesday's reversal, she tweeted: "Common sense prevailed."
The vote came following a new California law requiring city police forces to keep inventories of military-grade equipment and seek approval for their use.
After the vote, the board sent the issue to the committee for further review.
Robots are already present in police activities. To date, their use has been mostly restricted to intelligence and reconnaissance activities, such as assessing bombs or detecting threats in low-visibility environments.
However, there are some situations and locations in which the law allows for the arming of robots. For instance, in 2016, police in Dallas, Texas, used a robot armed with C-4 explosive to kill a sniper who had killed five officers and injured several more.
In contrast, in California, a new law that went into effect this year currently requires police and sheriff's departments to inventory military-grade equipment and seek approval for their use.
This new requirement has prompted the San Francisco vote, as well as a similar one in Oakland, where police departments also backed down from their proposal to arm robots with shotguns following public opposition.
Despite the U-turn, San Francisco's new policy does allow the SFPD to use robots for situational awareness, such as sending the equipment into dangerous situations while officers stay behind.
"Having robots that have eyes and ears and can remove bombs, which happens from time to time, is something that we want the police department to do while we continue to have this very controversial discussion," said Aaron Peskin, a SFPD supervisor.
Other countries have also been joining on the trend of using robots for police activities. In 2017, the Dubai police department tested a humanoid robot capable of searching for known criminals and expressed its intention to have autonomous robots comprising 25 per cent of its patrolling force by 2030.
