“Yes, this is 1st use of robot in this way in policing,” said a tweet from New America Foundation fellow Peter Singer, who heads a technology advisory firm and has written on 21st century warfare.

Singer noted that one device known as the MARCbot “has been ad hoc used this way by troops in Iraq”.

Dallas authorities did not offer details on the device, but the city’s emergency management inventory lists a Northrop Grumman Andros robot designed for bomb squads and the military, and some media reports said this could have been used in the killing.

Northrop’s website says the robot is “designed to defeat a wide range of threats including vehicle-borne improvised explosive devices”.

Matt Blaze, a University of Pennsylvania professor of computer and information science, said on Twitter that “stakes go *way* up when something like this is re-purposed as a weapon.”

Regardless of the methods in Dallas, the use of robots is expected to grow, to handle potentially dangerous missions in law enforcement and the military.

One robot developed by China’s National Defence University called AnBot has been designed for “an important role in enhancing the country’s anti-terrorism and anti-riot measures,” according to its website.

A Popular Science article authored by Singer with Jeffrey Lin, says AnBot is among several machines designed for police patrolling.

“AnBot’s most controversial feature is naturally the on-board ‘electrically charged riot control tool’ [likely a Taser or extendable cattle prod]. This can only be fired by the AnBot’s human remote operators,” they wrote in the April article.

“The AnBot’s large size means that it has room to mount other law enforcement gear, like tear gas canisters and other less lethal weaponry.”

Another robot developed by California startup Knightscope uses “advanced anomaly detection,” and according to the website TechCrunch is being used to augment security at some Silicon Valley tech firms and a shopping mall.

Researchers at Florida International University meanwhile have been working on a TeleBot that would allow disabled police officers to control a humanoid robot.

The robot, described in some reports as similar to the RoboCop in films from 1987 and 2014, was designed “to look intimidating and authoritative enough for citizens to obey the commands,” but with a “friendly appearance” that also makes it approachable, according to a research paper.

The strategy opens a new chapter in the escalating use of remote and semi-autonomous devices to fight crime and protect lives. It also raises new questions over when it’s appropriate to dispatch a robot to kill dangerous suspects instead of continuing to negotiate their surrender.

“If lethally equipped robots can be used in this situation, when else can they be used?” says Elizabeth Joh, a University of California at Davis law professor who has followed US law enforcement’s use of technology. “Extreme emergencies shouldn’t define the scope of more ordinary situations where police may want to use robots that are capable of harm.”

Dallas Police Chief David Brown defended his department’s decision. “Other options would have exposed our officers to great danger,” he said.

Dallas Mayor Mike Rawlings applauded Brown for making “the right call” and said he would have no qualms about resorting to the same strategy in the future. “When there’s no other way, I think this is a good example,” he said. “The key thing is to keep our police out of harm’s way.”

A 2014 report by Human Rights Watch and Harvard Law School’s International Human Rights Clinic raised concerns about the use of fully autonomous weapons in law enforcement operations.

“They could not be preprogrammed to handle all law enforcement scenarios,” the report said.

“And they would lack human qualities, such as judgment and empathy, that enable police to avoid unlawful arbitrary killings in unforeseen situations.”

Additional reporting by Associated Press