San Francisco has approved a policy that could let police use robots to kill in an emergency. The Board of Supervisors passed the measure 8-3 after a debate lasting more than two hours, setting off questions about when a robot may be used as force and how narrow that permission will remain.
What the policy allows
The language appears in a new Law Enforcement Equipment Policy submitted by the San Francisco Police Department (SFPD). The department filed the document in response to California Assembly Bill 481, which requires law enforcement agencies to provide a written inventory of military equipment they use.
That inventory includes an armored Lenco BearCat vehicle, flash-bang grenades and 15 submachine guns. It also lists 17 robots owned by the SFPD, 12 of which are fully functioning. None were built specifically to kill; most are used to find and dispose of bombs.
The approved language could allow those robots, or other robots, to kill when the risk of death to members of the public or officers is imminent and outweighs other force options available to the department. It does not mean the listed robots were designed for that role. It creates a possibility for their use in a specific kind of emergency.
Safety and the limits of deadly force
The debate turns in part on how existing ideas about deadly force apply when the equipment is a robot. Police in the U.S. are authorized to shoot in circumstances that include self-defense or situations where someone faces death or serious bodily harm. A robot is not a gun, though robots equipped with guns exist, and the vote effectively permits weaponizing robots in cases of this kind.
SFPD spokesperson Allison Maxie said robots equipped this way would be used only in extreme circumstances to save lives or prevent further loss of innocent lives. Maxie also said explosives could be used to contact, incapacitate or disorient a violent, armed or dangerous suspect.
That explanation describes a limited purpose: stopping an immediate threat when other options are inadequate. The policy’s critics are concerned with how that permission might work in practice, including who decides that the situation meets the threshold and what prevents the tool from being used in other circumstances.
Why critics warn about mission creep
The concern is that equipment introduced for an exceptional emergency can move into ordinary policing. The Electronic Frontier Foundation warned about “mission creep” in the use of armed robots, arguing that technologies first authorized for extreme situations have later appeared at protests or in responses to minor crimes.
The organization pointed to cell site simulators, also called “Stingrays,” which were developed for foreign battlefields and later used by law enforcement to catch immigrants and a man who stole $57 worth of food. It also cited police use of face surveillance and Amazon Ring doorbell cameras against BLM protesters.
The SFPD has no current plans to put guns on robots, Maxie said. But the policy’s approval comes amid broader growth in robots developed for military uses, and the article notes that the language could make it easier to acquire systems already designed as weapons.
There is a prior example of a robot used in a fatal police operation. In July 2016, the Dallas Police Department killed a suspect using a robot carrying a bomb, in what is believed to be the first such case in U.S. history. Police chief David Brown said officers saw no other option but to use the bomb robot and place a device on its extension for detonation where the suspect was.
A contested decision in San Francisco
The vote drew attention because it appeared to conflict with San Francisco’s image as one of America’s most liberal cities. Board member Rafael Mandelman argued that progressive politicians worry about appearing hostile to police, saying that perception could hurt progressives, the Board of Supervisors and Democrats nationally.
Rules Committee chair Aaron Peskin had tried to add language stating, “Robots shall not be used as a Use of Force against any person.” The SFPD reportedly crossed out that line.
San Francisco’s debate followed a different outcome in Oakland. After public backlash, the Oakland Police Department said it would not add armed remote vehicles and would stop exploring that option, despite earlier committee discussions with the Oakland Police Commission and community members.
Board President Shamann Walton warned that expanding police weaponry could increase opportunities for harmful interactions between police and people of color. The opposing views leave a central question unresolved: how a policy intended for emergencies can ensure that robot use remains limited to those circumstances.