This is Not the Killer Robot You’re Looking For: Dallas Police Used a Precision-Guided Munition to Kill the Shooter

July 11, 2016

For special access to experts and other members of the national security community, check out the new War on the Rocks membership.

A standoff between police and one of the suspects in a Dallas shooting spree, which left at least five police officers dead and seven others wounded Thursday night, ended after the suspect was killed when a robot delivered and detonated explosives where he was holed up, according to local law enforcement officials.

The move represents a potentially unprecedented use of robots to deliberately deliver lethal force in domestic policing, according to experts, raising questions about how local law enforcement officials are deploying the high-tech tools that increasingly fill their arsenals.

The Washington Post, July 8, 2016


I knew as soon as I saw the headline that the “killer robot” articles would start. That the use of a compact remote-controlled vehicle to selectively eliminate a dangerous, armed killer in a protected position would cause science and technology writers everywhere to collectively gasp and head to their keyboards. There, they would engage in a spinning whirlwind of predictive doom, calling for new regulations, stoking fears of hordes of government-controlled killer robots, and speculating on the future of civilization. But all the hyperventilating over this by technologists obscures the fact that robotic devices have been used to deliver deadly explosives for decades — almost 100 years, if you count the Kettering Bug. Rather than focusing on the robotic delivery of the explosive, it is more useful to understand this as the directed application of a precision-guided munition (PGM) under conditions that clearly called for one.

The facts of the event are generally agreed upon in the press. A domestic terror attack began at roughly 2100 hours, when Micah Xavier Johnson opened fire on police officers at the scene of a peaceful protest in Dallas. An exchange of fire began a chaotic firefight in which the assailant changed positions several times before holing up on the second floor of El Centro College’s C Building. There, the shooter established contact with police, expressed his anger, and indicated that he had placed explosive devices downtown. Fire was again exchanged, and the “negotiations” broke down. Dallas police attached a small explosive device to an explosive ordnance disposal unit, emplaced the device adjacent to the shooter’s position, and detonated it, killing the shooter without further danger to police or the public.

While this may have been a new tactic for police in the United States, it is hardly a new tactic. Wehrmacht soldiers in World War II employed the SdKfz. 302 Goliath, a remote-controlled mobile mine that looked like a subscale World War I tank and carried a substantial explosive charge. The weapon trailed a 2,100-foot cable and was steered by a joystick on a simple control box, held by a soldier (presumably) under cover. The Goliath itself was inspired by a French device captured in 1940. The Goliath was not a large success: cutting the cable rendered it ineffective, it moved at a mere 6 miles per hour, and it was vulnerable to counter-fire. Nevertheless, over 7,500 were produced, and they were used on every front.

On Christmas Day 2007, I went on patrol through Baghdad with my favorite combat engineers, the 1138th Sapper Company of the Missouri National Guard. It was an unusually challenging and memorable patrol, capped off by the discovery on the way home of a surface-laid improvised explosive device (IED) on the median of Route Irish, one of the main roads to the Baghdad Airport. We dismounted, cleared the area (including the airspace), and sent our little remote-controlled Talon robot up to the device with two pounds of C4 explosive clutched firmly in its robotic claw. The Talon is a rugged tracked robot with a variety of uses beyond military employment. If it were used by an 11-year-old in a public park to pick up trash, it would be just a toy, albeit an expensive one.

Controlling the Talon from inside our massive Buffalo Mine-Resistant Ambush Protected Vehicle (MRAP), the operator moved the robot into position, placed the C4 on the IED, and moved off. As the new guy, I pulled the detonator ring and … nothing happened. Having all relearned the important lesson that one should not close the MRAP’s armored door on the detonator cord, we robotically emplaced another block of C4 and pulled the ring again, successfully. No collateral damage, no injuries, and little fuss. I mention this little event not because it was exceptional, but rather because it wasn’t. It was a routine emplacement of an explosive charge to a precise location for a planned application of minimal force to achieve an effect. Our robot was a precision-guidance package for delivering an explosive weapon to the target.

That’s what the Dallas Police Department did. Faced with a poor tactical situation in which a number of officers were known dead and wounded, the department realized it was dealing with a dangerous individual who claimed to be in possession of explosive devices. So, having localized the target and assessed the collateral damage potential, Dallas police officers selectively employed a precision weapon to eliminate the threat. This is not just a semantic argument — it is exactly what they did. Yes, they used a robot. So what? We use robots to assemble cars, survey terrain, photograph weddings, sort inventory, and drill precise holes.  Robots are merely tools, and it’s hardly necessary to fly into a tizzy when someone creatively uses an old tool in a new way. Humans are tool-users, after all.

Unfortunately for them, police departments are not routinely equipped with precision weapons. Fortunately for them, they had the vision and flexibility to build and employ one on the fly, thereby accomplishing everything we demand from a police department under very trying circumstances. It is very easy, when taking casualties, to lose some element of discipline or control and do far more damage to the surroundings than necessary to contain the threat. The Dallas Police Department did not set loose a killer robot. They emplaced a charge using a precision method that posed the least risk to their force. The Dallas police should be commended for their restraint, discipline, and ingenuity in the face of chaos, confusion, and death. Essentially, they used a PGM under conditions wherein that was exactly the proper response. My hat’s off to them.


Col. Mike “Starbaby” Pietrucha was an instructor electronic warfare officer in the F-4G Wild Weasel and the F-15E Strike Eagle, amassing 156 combat missions and taking part in 2.5 SAM kills over 10 combat deployments. As an irregular warfare operations officer, Colonel Pietrucha has two additional combat deployments in the company of U.S. Army infantry, combat engineer, and military police units in Iraq and Afghanistan. The views expressed are those of the author and do not necessarily reflect the official policy or position of the Department of the Air Force or any part of the U.S. government.

We have retired our comments section, but if you want to talk to other members of the natsec community about War on the Rocks articles, the War Hall is the place for you. Check out our membership at!

15 thoughts on “This is Not the Killer Robot You’re Looking For: Dallas Police Used a Precision-Guided Munition to Kill the Shooter

    1. Extra-legal? An armed criminal was killed while committing murder, and declaring intent to continue. No death squads, black helicopters, or killer robots in sight.

    2. So law enforcement use of the robot was somehow extra-legal? Nope, entirely as legal as using their own sniper force or sidearms. The shooter was not legal and using all available means of force on his part to kill as many LE officers as he could. Stopping him and making him no longer a threat or committing illegal acts was entirely legal.

  1. With all due respect to the Colonel (as a fellow military aviator I have enjoyed his many articles on airpower), I think he is treading into dangerous territory with this article. I would like to refer the Colonel to a couple of articles he may want to go back and read.

    Certainly the “legality” of using a remote Controlled Vehicle (RCV) to deliver an explosive device as a means of “neutralizing and/or killing an active threat” is something that we should, as a society, be debating. Personally, I feel this becomes a slippery slope and one that could have far reaching legal and ethical implications for our Constitutional Republic.

    For any student of military history, the Colonel is correct in his assessment of MILITARIES using this type of technology for around a century now. That shouldn’t come as a surprise to anyone. So in arguing the use of this technology in a law enforcement setting, one finds it curious that the example(s) provided in this article did not include one single case where this technology was employed by U.S. civilian LE as an accepted practice. Using this logic, would it not make just as much sense to start providing law enforcement departments all over the country with heavy armor (tanks) and remotely piloted aircraft (RPA) armed with PGMs to neutralize and/or kill future individuals who are engaging in deadly violent acts against the local community?

    One last parting shot, I also spent a little time in the sandbox just like the Colonel. I was present for Operation Phantom Fury (the second battle of Fallujah) and at least a month prior to the operation the U.S. and allies started disseminating warnings to the city’s inhabitants that a large scale assault was about to take place and to “vacate immediately”. It is estimated that 70-90% of the cities inhabitants (civilians) fled prior to the operation commencing. Considering the U.S. has the largest and most sophisticated arsenal of attack aircraft and PGMs in the world, why didn’t we just bomb the every living daylights out of the city ( a la World War II heavy bombing strategy type stuff) and save precious allied lives by foregoing sending in ground forces? In fact, one could broaden this concept to all current U.S. military operations around the globe. Why risk American lives when we could literally pound “terrorists” with all our fancy high tech weaponry. I mean, if law enforcement can start using this logic, then shouldn’t we as a society say the same goes for our military when they are operating in foreign countries where there’s no fear of killing and/or injuring any U.S. citizens?

    For the record, I have many friends that have been in or are LE. There’s no doubt it’s a hard, dangerous, and thankless job at times. However, I’m not sure I’m on-board this idea of using military TTPs in a civilian LE setting. More importantly, if this is the road we go down as a society I think the opposite effect will happen. It will create more disobedience, distrust with LE, and possibly an increase in mass violence. One only needs to look in the mirror of the last 15+ years of U.S. military operations around the globe to see that what the U.S. has done hasn’t exactly made it a safer place to live (or be an American).

    1. A few things to consider:

      The draw-down of the occupation of Iraq did create a second-order market for military equipment to LE.

      Many EOD robots that LE has now came from the military draw-down. And, you should know the SWAT teams do have Snipers too. It did get out of hand and much of the more militarized equipment has since been taken back.

      And the CBP does fly MQ-9s.

      The distinction is one is innocent until proven guilty but balanced with public safety. Before you mention the legalities with innocent until proven guilty, realize that if the EOD robot had been a person with a concealed carry permit and neutralized the shooter that person would not be arrested. If it were a K9 that attacked the shooter would we be having a conversation about the merits of police dogs to mitigate risk to human lives?

      Taking this example airborne changes the dynamics to a point that there is no comparison to make.

      As for your parting shot, I was also there for Phantom Fury (cue in Gunny Highway’s comment about chewing the same dirt). And, I am also familiar with what happened in Kobani, Syria in late 2014 too. What you proposed for Fallujah is pretty much what happened in Kobani. Literally thousands of PGMs dropped (at the price of a new car each), destroying a city with precision, one building at a time; street by street, block by block. The city was ‘saved’, though it is a shell of what it used to be.

      1. Greetings TBN,

        Just to be clear, I’m not comparing myself to Gunny Highway. I’d need to be taller and have a lot more charisma to fulfill that role! I am aware that after some negative press was brought forth several years ago about “military grade equipment” being handed down to local LE that many of these items were “returned” to DoD. That wasn’t my primary concern though.

        My concern is “To what extent should LE be armed and what TTPs will be acceptable, legal, and ethical?” Law enforcement should be there to “keep the peace” (that’s why they are called officers of the peace correct?) and to protect life (the citizens they serve first, themselves second). I tried to tie this into the fact that the U.S. military has been operating in hostile environments for a very long time and often with very restrictive and sometimes confusing ROE. We do this to ensure that we don’t kill “innocent” people (non-combatants), often at the risk of our own personnel. That’s why I am confused at how we, as a society, can accept the use of a “robot” with an explosive device strapped to it to be considered an acceptable, legal, and ethical tactic for dealing with a violent threat. As you pointed out, I’m sure Dallas has a SWAT team, isn’t this what they get paid (and train) to do? No one wants to lose life when it isn’t necessary, but using a VBIED (in essence) in a Constitutional Republic against a U.S. citizen (no matter what act they are committing) seems like a huge stretch in everything we, as a society, should stand for.

        I also completely understand that aviation brings in a new set of dilemmas, but that is exactly where I was going with my first comment. If we allow LE to use a VBIED against suspects, then how long till this morphs into using air assets for the locating, identifying, and targeting of individuals (suspects) on U.S. soil? As you are well aware, the U.S. military has killed many non-combatants “by accident” and the local blowback certainly didn’t win us many friends. In fact, I’m sure we’ve made a lot more enemies than friends the last 15+ years. What happens when a local LE operation goes awry and instead of killing the bad guy, we blow up a family of 5 “by accident”? What about due process?

        I’m not sure where you were going with your tie in of Kobani? It seems you were validating my point of, destruction is easy but creating peace is hard?

        Here’s an interesting link on the use of deadly force by LE. This study was conducted by the DoJ. Look at the date it was released (1979) and then read what the study/report says! If you didn’t know the date, you’d think we were reading something from today. So much for progress eh?

    2. Carpet bombing? Seriously? We in no way have the capability to do that having a very limited bomber fleet that is not even 10% of what we had in WW II. Indiscriminate bomb simply kills more civilians and not the enemy for the most part.

      And claiming LE should do the same is missing the point; our military is a reactive force that uses its overmatch to deter attacks and then, when attacks occur, moves out and engages. LE is not an offensive force it is a protective force using extreme methods of taking out the criminal(s) only as required; hence why there are high speed car chases and hostage negotiations rather than missile strikes on the offending vehicle and collapsing the building on the criminal and their hostages.

      As for using military TTPs in a civilian setting, concur. Yet there is only so many ways to do something–a robot delivered explosive is a robot delivered explosive; how, exactly, is it to be done differently?

      1. Patrick,

        I know my reply is a little on the late side, but I always enjoy a good debate. To clarify, I was never advocating the tactic of carpet bombing in my analogy. I was just comparing the fact that the U.S., during combat operations in Iraq during the 2004/2005 period, had complete control of the skies and a highly sophisticated arsenal of attack aircraft at its disposal.

        However, since you claimed that our modern Air Force could not compete in quantities that the Army Air Forces of the World War II era enjoyed, let’s do a quick breakdown using some extrapolated numbers.

        In 1943, the 8th Air Force operating out of England was quickly building up its operational bomb groups as the U.S. was ratcheting up its daytime heavy bombing raids of German logistics, military targets, and production facilities. A typical B-17 bomb group was comprised of four squadrons and each squadron could have anywhere between six to thirteen aircraft assigned to it depending on time period (as the time passed U.S. manufacturing was able to churn out greater numbers), aircraft availability, combat losses, and available crews (combat losses were quite horrific well into 1943; also, flight schools had to produce a large number of qualified men to operate these machines) to man the aircraft. Boeing, who designed the aircraft as you probably are aware, developed the B-17 to carry a 4,000 lb. bomb load.

        A modern heavy bomber wing in the Air Force has three bombardment squadrons, with each squadron having fifteen B-52 aircraft assigned to it (according to my research). According to Boeing, who also designed the B-52, the aircraft was developed to carry 70,000 lb. bomb load. As I am sure you are aware, under normal aircraft operations, it is extremely rare to have a full compliment of aircraft operational at the same time for a given mission.

        So for our comparison, lets use some reasonable assumptions in our numbers (keeping in mind that I highly doubt modern Air Force doctrine calls for the use of heavy bomber squadrons to drop “dumb” bombs en masse like its predecessors did during the heavy bombing campaigns of World War II).

        Using a typical World War II heavy bomb group of four squadrons, let’s assume each squadron is able to put ten bombers into the air for this mission. Each bomber is carrying a full compliment of M64 General Purpose 500 lb. bombs for a total of 4,000 lbs. per aircraft. With forty aircraft, that’s 160,000 lbs. of ordnance or 320 bombs.

        For the modern heavy bomber comparison, we will use just ONE squadron of B-52s. Trying to be fair, we will use ten aircraft launching, each carrying a full compliment of MK 82 General Purpose 500lb. bombs. That would mean each aircraft is carrying 140 bombs for a total of 70,000 lbs.! With ten aircraft, you are carrying 700,000 lbs. of ordnance or 1,400 bombs. That’s over four times as many bombs using 1/4 the aircraft for the same mission! Thus, I think it is safe to argue that while the overall number of aircraft available is less than its bomber counterparts of World War II, the capability is much greater per aircraft which equates to a lot more ordnance being dropped.

        I know we have gone of the rails here with this comparison, but my earlier claim is still valid. The U.S. Air Force is quite capable of generating levels of ordnance that exceed its 1940s counterparts using a lot less airframes to do so.

        My original question remains, if the U.S. military was to adopt the attitude of “ensuring the safety of its service members is priority number one” then why didn’t we just pound Fallujah into the sand using all our expensive aircraft instead of sending in the grunts? It certainly would have saved a lot of U.S. lives correct? I think we should be debating exactly what is acceptable tactics for CIVILIAN law enforcement departments to be utilizing in this country. As another commenter has stated since I posted my original comments, should we accept police departments to become judge, jury, and executioner in these situations? There’s also the consideration of proportionality. Granted this may be a military concept under the laws of warfare, but why should law enforcement get a more lax set of guidelines to follow compared to the military?

  2. I’m thinking that while well written, the article suffers from a “forest for the trees’ attribute specific to the fact that this is a first use of a armed unmanned ground vehicle (UGV) under control of law enforcement in a domestic setting and not the near routine use of such UGVs or UAS downrange for military purposes.

    Claiming “robots are merely tools’ means a discussion has to occur because using tools against humans is a major issue outside of the military, law enforcement, and home protection against walkers. IOW, the Colonel is coming at this from a standpoint of being military and not being civilian.

    Having spent time downrange before the advent of more than a handful of UAS the “overtrust” aspect is looming related to technology specific to unmanned systems and their potential once more and more autonomy aspects are added. Not saying we need to worry about lightning strikes changing how a robot’s autonomy or even artificial intelligence is ‘redirected’ as it was with “Number 5” or the combat drone in the movie “Stealth.” But having been embedded with UAS (primarily) for the last 15 years via policy, programmatics, fielding, and concept development here and in NATO, the phrase from “Ghostbusters 2” comes to mind when Bill Murray notes at one point “boys! boys! you’re scaring the straights, okay?!”

    This is the aspect the article completely misses because it is not normal for the straights…sorry, civilians…to consider routine use of armed unmanned systems because they don’t see the robots that drill holes, make cars (unless they watch the Discovery Channel), etc. Not understanding mindsets gets us into more conflicts (verbal and otherwise) than should occur. As Sun Tzu said so long ago “know your adversary and yourself and you will win 100 battles…” paraphrasing.

    1. Pat — that’s exactly the point, though…as soon as someone uses a remote-controlled device (unfortunately, I can’t put that into italics), against a human, the anti-machine community pops up and starts gibbering about how we’ve crested the slope on the way to autonomous mechanical killers. These machines are not new to law enforcement — bomb squads have had them for years. If Dallas PD had used this to bring a cell phone (or lunch) to the guy, and a police sniper later killed him, this discussion wouldn’t be happening.

      This is another variation on the arguments regarding airstrikes with umanned vs manned aircraft. The real question is: did this situation satisfy the conditions for the police to use lethal force. If that’s a “yes,” then the police had a responsibility to minimize risk to the surrounding public and to themselves…which they did, and with more control that if they’d authorized a sniper.

  3. Delivering explosives by a remote device is technologically very simple and is not the surprising part of this case.

    What is unusual and worrying is that police (civilians) were justified in employing lethal force when no one was in clear danger. I realize that the standoff was virtually guaranteed to end with the suspect’s death, but the semantics of how he died are very important.

    It is crucially important that the police only be allowed to use lethal force to protect life, otherwise they become judge, jury, and executioner. To kill without a judge and jury is either self-defense or murder, so I really hope the police have a reason their bomb robot was self-defense.

    1. I think you understand my concern and you did a much better job of being succinct about the topic we should be debating. I agree, I have no issue with end result, it was the ways and means the DPD went about it that should be scrutinized. I feel like this has huge legal and ethical considerations for future policing and our society in general.

  4. Excellent article, as one would expect from Col Pietrucha. And one more example why the category ‘Killer Robot’ is an analytically useless NatSec ‘Folk Devil.’

    As he points out with the Kettering Bug, there is and long has been a continuity and a spectrum between PGMs and autopilot-equipped-aircraft. Understanding the strengths and weaknesses of the poles of that spectrum is more useful than a moral panic that confuses means and ends.