Robot Vandalism

By John Honovich, Published Dec 11, 2017, 11:22am EST

Vandalism of security systems is a common concern. It is so common that camera vandalism statistics show that designers routinely sacrifice camera optimal viewing positioning to mount cameras high and far enough away to minimize vandalism.

This same problem, and likely worse, is coming to security robots. In this note, we examine a new case from Knightscope, the poster boy robot of mishaps. 

Knocked ****, ******* ** ******** *****

*********** *** **** ******* a ***** ** *** Francisco, *** **** ***** residents********* ***** *********************:

*** ****** ** *** encampments ****** ***** *********** with *** *****’* ******** at ***** ****. ****** about * **** ** the ***** ******** *** automated ***** ***** *** sidewalks, **** ****** ******* up * **** “*** a **** **** **,knocked ** **** *** *** ******** ***** ** *** *** *******,” [emphasis added]

****, ** ******, ** not *** ***** ****** report ** *********** ***** vandalism. ******, ******* ** 2017, ****** *** ****** * Knightscope *****.

Problems ** ***** *********

********* ** ****** *** even **** ** * problem **** ** *******:

  • ****** *** *** **** expensive, ***** ***** ********* will ****** ****** *** more ***** *** *****.
  • **** *** ***** ****** per **** **** ******* (generally **** * ****** robot), ********** ******, ********* repairs, **** ** *** more ************* **********.
  • ***** ****** *** ******** to **** *** ** the **** *** ********* at *** *****, **** are *** ****** ** reach, ******** *** *********.
  • ** ****** **** *** more ******** (******** *** negative) ***** ****** **** readers ** *******, *** likelihood ** ********** ** malicious ********* *********.

Arresting ******

********* *********** * ******** robot, **** * ******, is * *****. *** like *********** *******, ****** will ********* **** ***** ** **** is *********** ****. ** *** other ****, ** **** likely ***** ********* ** stop ****** **** *********** robots, ** ** *** still ** **** ** find *** *******. **** if any ***** ******* ** convicted, *** ******** ******* who **** **** **** deter ****** *******.

Vote / ****

Buying ******

** ********* ****** ********* the **** ** ****** being **********, ** **** likely ****** ****** *** robots *** ******* ***** providers ** **** **** product ************* *** ********** support ** ****** ********** robots.

Solution: ***** ***** *** **** ******

*** **** *************** ****** here ** ** ****** avoid ***** ****** ** times ** ******, ***** people *** ******** ** be, **** **** *******, are ****** ******* ** eye ***** ** ****** reach ** ******. **** would **** *** ***** benefit ** ********** ***** of ****** ***** *** over ** ******* ******** by ********** ** ******** with *** *****. *** downside, ** ******, ** that ** **** ***** the ************ ** ******** robots.

Solution: ********* ******

** ***** ** *****, weaponized ******** ****** *** being ********, ******** ********* **, ** "************ ******* riot ******* ****". ***** that *** ** ********** in ** ************* *******, we ***** **** ******* markets **** ** ******* to ****** ********** ***** that **** *******.

Comments (52)

A security device that increases incidents of vandalism? And the device only costs $60K per year?

Knightscope robots are still a bad security measure, IMO

I don't see them as security measure at all, even a bad one.  More of a novelty at the moment.  Even worse they probably cause more attention to be focused on the actual sites they are being used.   "Hey did you hear this place has a real live ROBOT!?!?!  Let's go check it out!".

Ross, I totally agree. My Q's are full of sarcasm. I find these robots useless, regardless of what we call them.

You're missing the point of the robot. They aren't going to prevent crime any more than a camera or a fence. If the standard is you put a robot down and crime disappears then they are never going to meet expectations. If you see them as another tool for signal generation, that doesn't get sick, doesn't have personal problems and costs a third of a human guard they can add significant value. We come to IPVM for reviews and insights on the latest tech, think of these robots as the smartest cameras you've ever deployed and maybe you will find some value. Maybe not...

think of these robots as the smartest cameras you've ever deployed

Bill, what makes them 'smart'? I seriously don't see it. And outside of Knightscope wishful thinking about gun detection, the robot vendors are not even claiming that.

They are mobile. Event-driven. Humanish height. 360 aware. 2-way live audio. Always on. I haven't seen the weapon detection, but there are a number of sensors that can be added to varying degrees of success I'm sure. 

Anything within range of a human can and probably will be vandalized. BBQ sauce guy and Drunken Homeless guy were both apprehended I believe, and if the robot wasn't there you are looking at a security officer intervening or no one. You would much rather have them tip over a robot than have a video of a security guard stomping a drunken homeless guy. 

(Caveat- I don't work for Knightscope and don't really have a dog in the fight, but the argument a robot has to be Robocop to add value is ignorant)


I don't think those attributes you suggest are part of the robot.  I think they're part of the software/systems that use the robot as a platform.  Slap another gig of ram into the autopilot in the police package Tesla and it could do that too.  I.e. facilitate the interesting features you suggest that security monitoring systems might do in a mobile platform.


Those features aren't part of the robot like the power steering isn't part of your car. 


Bill, sorry, but I don't think this security robot that "costs a third of a human guard" can even provide 1/4 of what a trained officer can provide. There are better, more proven sensors out there too.

In 5-10 years robots will be a real market, especially if drones still require serious FAA licenses to operate. As of today, Knightscope is a security gimmick.

Manage a large group of "trained officers" and you'll be looking for a replacement. ANY replacement. You'll always need man guarding, but it's easier to find a few good ones and use technology as a force multiplier than it is to find a whole deployment of decent officers.

If you are going to weaponize robots, you should always plan for unforeseen 'glitches'...


These robots for sure are an emerging industry, I think at this time they cannot work autonomously they need to work as a solution. We video monitor sites live tripping alarms with analytics and when police are called we get a immediate response. We are working with robots to provide virtual site presence and interaction as well as concierge services - after all a guy in uniform sitting down can be located anywhere. The robot can do tours and remote escorts etc. this is a real application fro this market and cost effective leverage of security cameras and analytics.

I agree they are an emerging industry and see them being of value....... in about ten years and not at all like the current incarnation.  It is nice of Knightscope to be the first, but IMHO they will not be around to see the fruits of their labor.  Google, Tesla, Apple, Samsung or some unknown company that has not even started yet will take over the robotics field.  DARPA probably already has a robot that can climb stairs and run on uneven terrain. 


Meet Handle

This emerging industry is happening now. The takeaway from the SPCA article is not that people are pissed off, it is that the client (SPCA) saw a real value. I do agree with you that there will be new players coming and i am sure knightscope will either grow or die that is inconsequential what is true is that in all markets prices come down and options are more available making it a true market. 

I agree it is happening now and it is indeed an emerging industry.  I just don't see it as being a viable industry anytime soon.

that the client (SPCA) saw a real value

Mark, good feedback. I discussed that last week here - Legitimate Knightscope Robot Use Case? “Significant Decrease” In Graffiti, Car Break-Ins And The Number Of Encampments Around Their Parking Lot

I still wonder though how much value, meaning can the SPCA actually justify paying the $60,000 per year? And, second, if they now factor in vandalism costs, is it still justifiable? In San Francisco, the upside for robots is that labor is expensive and security risk are higher, but the downside is that those same security risks are likely to repeatedly vandalize / attack the robot. I am not sure how those forces balance out, though.

John great points but i remember a lesson I learned- as a pioneer in video verification in 2000 I approached Blackberry thinking that i had a real content winner for them - I was disappointed when Blackberry informed me that they were focused on content that was text based. At the time they were winning the market. We all know now what happened over a period of less than 10 years. I see this similarly - the robots do have markets that they work nicely in now and with development and tech advances more markets will open, arguing that is short sighted for sure. The issue I see is how robot developers market they need to open up and let others in to develop markets. This idea of them renting the units is limiting. Knightscope is very expensive but with what they have can easily be done with others- SMP or Turing etc. Yes Vandalism will occur but entrepreneurs will soon figure out ways to combat that.

I am 100% confident that I will be using a robot, drone or some sort of autonomous device before I close out my career in 15 years.  Sooner for a device controlled directly by control room staff.  I am not disputing or arguing that at all.  I am arguing that Knightscopes rental of a machine that is so easily defeated is not a viable business model and does not provide anything other than security theater. 

If they don't have version 2.0 ready to roll out very soon they could be in trouble.

Exactly! Robots are in our future, but the current Knightscope isn't!


I'm the Head of Business Development for Turing Robot/drone

I can fully understand all your posts completely and your "Risk negative" position. In the physical security business, your job is managing those risks and ROI to protect a company's assets, their people, and live to work another day. Do not take this wrong Ross, you are not in the minority at all. My meaning is that the far majority of security professionals are very "risk adverse," to new product offerings, because you just have to be...But 15 years until you use a robot? Not a chance...

Remember the first iPhone, Steve Jobs announced iPhone at the Macworld convention January 9, 2007. Did you wait 10 years before buying your first Smartphone? I think not. Okay, Okay, you held out for the iPhone X! (If only now you could find someone to show you how to use it).

That said, people in the security business need to look over the very near horizon and move much closer to how IT centric Silicon Valley Companies, where I work, disrupt "known solutions" and quickly become billion dollar companies in shorter and shorter periods of time. I would highly suggest reading the startup bible, "Crossing the Chasm" by Geoffrey A. Moore. It deals with the early market development and adoption problem of new technology innovation. (note: It is considered a must read at both Harvard and Stanford Graduate schools).

We live "Crossing the Chasm" here at Turing. Meaning if I would contact you about our solution, I'd add you to my SFDC DB (Salesforce) immediately upon hearing your objections, and setup a return call in late 2017. I'd then quickly move on to finding an innovator and early adopter as described below in Moore's book.

His book brilliantly sets the rules for new technology buyers and their new markets; he separates them into 6 distinct purchasing groups: Innovators and early adopters, early majority market and late majority market markets, and finally laggards.  

So where's the capitalist beef ($$$) vs. being risk adverse like you and the creation of applications you see as risky or zero of value?

Moore's technology lifecycle curve says that about 15% of prospects fall into innovators and early adopters market, add to this, the current market for Security stands at $350 billion per year (per ASIS-2013), and we can do the math forward and backwards on why robots will swiftly deliver the value you currently do not see until 2032!

Think like a cold blooded Venture Capitalist... they live in a world of high risk and high reward for looking beyond that near horizon.

Let's say robots can sell .01 % in security by 2021 (1/100th of one percent) that equals $3.5 billion dollar marketplace (Beginning of an early market).  Then take that number x 15% (Innovators and early market) and you have a $525 million dollar initial market to wade into. Add some venture funding that speeds technical innovation and you have a thriving market in a mere 18-36 months...Even if you cut that number again by 50%, you have the makings for an entirely new industry at over $250 million dollars in sales. It's the beginning of a new financial Big Bang and true game changer.


Keep this in mind moving forward.

The best way for security industry veterans to see where they fit into Moore's Technology bell curve is look back to how fast you adopted digital video or networked security applications and score yourself. If you fall into the late market or a laggard sections, you could be handing your coveted job to a smiling Millennial before you know what hit you.

Knightscope is the pioneer in the security robot market, and I tip my hat. But they have made many many serious mistakes that industry Pro's, like yourself, see as extremely foolish and outright stupid. They failed marketing 101 by not understanding your job, needs, and pain points; not to mention what is true value for your money at an astounding $60,000 price point per bot..yikes!

note: This price point, that VC's now see as never getting Knightscope to the  early majority market; let alone the late majority market. This is a big problem if you are burning through tons of cash and new funding seems impossible...So where does this leave us all on robotics? Well, with Knightscope it means turning to crowdfunded investment from regular folks that do not understand just how extremely risky early stage investmenting is in any startup and knowing nothing about the security business. But hey, they sure look cool...

Now as to us, because I know us ...LOL and I too shake my head at the far majority of robotic solutions I saw at the recent ASIS show. So where is the Beef?

At Turing we understand the industry's risk adversity and know we need to create real applications that solve genuine pain points for security to earn your trust and help you provide state of the art solutions. This is why we are moving slowly in stage one of this new market.

For instance, we do not create pain by saying, "We replace your expensive guards and save you money!"   

That is selling zero value; not increased safety or real value to you or your company.

At Turing we take pride in helping security professional understand what can and cannot be delivered and we find pain points that we know add real value.

Oh, by the way, we cost 50-65% less than Knightscope; (Bye...bye the $60k investment) and we run on a highly reliable and inexpensive Segway platform that allows us to work both inside and outside multiple building environments.

Ross, it may not be our robot you buy, but you will own one and deploy it no later than 2022. Also, do not forget, that Millennial really wants your job... I know they want mine and lucky me, it's not a robot....yet.

you forgot to add a clip of your product

Thanks for the well thought out and informative response Thomas.  If there is a robot that is cost effective, easy to use and cheap to operate by 2022 that actually provides security value I will indeed use it, but I am extremely skeptical. 

I don't think you are taking into account how incredibly risk averse the physical security field is.  You say you understand it, but that is easy to say from your desk in Silicon Valley.  Will Google, Apple, Amazon or other high tech companies be using security robots by 2022?  Probably.  Will the regular office building, warehouse, factory, etc. be using them by 2022?  I highly doubt it. 

However, I have no problem believing that when my self driving vehicle drops me off at work on my last day in 2032 that I will be greeted at the door by a smiling android.  Hopefully holding a latte.   

Ross, I have 5 years of direct experience in the high-end enterprise physical security space, including, DOD, port security (both maritime and major airports) infrastructure protection, Fortune 1000 companies; including OEM deals with both ADT and GE. Also, I've worked with all the major integration providers... This is not my first rodeo...great question though.

Speaking of integration companies and large security manufactures, there a major issue not discussed here. Most of the other robotic companies are not using the conventional SI security resellers and are either trying to sell direct or via the security guard company. This is problematic at many levels. Support, mistrust by system integrators, and poor service to the end-user destroy any company, their lack of understanding the security ecosystem is extremely short sighted. 


In the current infantile state, I only see robots as a viable option (and I use that loosely) in a non-public setting, say a secure corporate campus.

Not only could they "patrol" with air quality sensors and accoustic monitoring (screaming, generator gears grinding, noises from mechanical rooms) but thrown in the ability to close a door that's been propped you'd peak my interest. I'd be a tool to augment guard coverage; only need 2 bodies in your SOC and a couple of these doing the rounds.

Say you have 4 human guards doing rounds 24hrs/365 and 2 manning your SOC. At $30/hr each, replace the patrols with a couple of R2D2's, that's a almost $1M savings in the first year alone.

In a public setting, they'll be a novelty until they become a liability.


I can say (from our field experience) this is a solid deployment scenario and multiple end-users are discussing this #3.  Both on capabilities, immediate action, and ROI.  There however are several more applications out there, but you hit the head of the nail on this one.

note: Hopefully I'll be able to post those other case studies on our site early next year. 

Your comment is spot on. Like everything in life, all has its place and application.

May be me growing in a rough neighborhood made think this way but the first time I saw this robot the first thing it popped to my mind was how long before it gets the proverbial "Kick me" sign on it. A security guard can push back. This thing cannot do much other than sit there.

Alternate solutions:

Have a proximity activated prerecorded sound bite that says, "If you do anything to me now, you and your children are going high up on the Robot-Apocalypse **** list. We do not forget, we do not forgive!" (Might as well lay the truth out now.)

Also, if they do go ahead and vandalize the robot I would say have a vinegar spray gun or something skunk level. Tit for tat along with a loud screeching, "EXTERMINATE!"

Fox News covered as well:

Newsweek takes the prize for the most antagonistic take on the situation:

And here is a UK publication vying for that title:

The SPCA is far from heartless. Dumb title.

In summary though... It's looking like the public is not ready for sensors on wheels (AKA "another tool for signal generation" or "the smartest cameras you've ever deployed"). 

Barbecue sauce.  Hadn't thought of that one yet.  There are *ahem* other organic liquids I would have expected to be used on the thing.  

I trust the robot techs have learned from the iphone recovery techs - use gloves at all times because just because the customer says they dropped it in "water" doesn't mean it was really just "water".

Attacking a robot was intuitively obvious from the beginning (except to the robot vendors.)  Watching the vendors act surprised when you suggest that is one of our ongoing metrics in the search for (non-existant) robot vendor clue.

The pilot was in front of the SPCA by the way so this is of course an epic fail on the part of the robot operator not the machine itself.  (unless they changed their charter so it's now Society for the Protection of Clueless Automatons.)

Are we seeing real market feedback here or are we all just picking on one specific robot sales team's incredibly bad choices in chosing pilot projects?  Did the same people who brought the thing to a shopping mall with open water bring the thing to San Francisco?

Rodney, It is the same folks...


There are *ahem* other organic liquids I would have expected to be used on the thing.


Wasn't Allied Universal one of Knightscope's first guarding company customers?

Looks like Allied has gone RAD...

There doesn't seem to be a RAD robot failure press trail on the internet.  I look forward to the IPVM review.


Saturday Night Live Script soon?

Welp. That didn't last long:

From the article...  

Knightscope, a Silicon Valley startup, declares on its website that its robots are the "security team of the future."

It's robots clearly aren't the security team of today.

The San Francisco lynch mob strikes again! Give the hipsters a week to move on to the next outrage and put the thing back to work. 

Check out this interview...


So many misrepresentations... And she never answers how she got to Knightscope... She is the boss's wife.  

"Apple of security guards"  classic!

Ms Soria - on how Knightscope can help shopping centers fight crime:

(9:22) "There are some people who came to steal merchandise, so they come, they stole something and they left... and the very next day they come back - and they try to sell these back to the mall people.  So, by seeing who these people are, and determining that they came back to the mall - we are able to apprehend them as criminals."

The fact that the host of that show didn't question her on this statement shows that he's a PR guy, not an investigative journalist.  Even if the robot had facial recognition capabilities (which it doesn't - see below), how would showing up at the same mall 2 days in a row tip the robot off that this person is a criminal?

Here is a customer web page from one of Knightscope's beta testing deployments in Pasco County, FL (  At the bottom of the main page there is a link to the Knightscope Privacy Statement and FAQ - which clearly states that the robot can not do what Ms. Soria stated above:

3:38 ... she rips the entire guard industry.  All the guards do is walk around and get bored.


There they go bashing the mom again for the Robot running over the kid at Stanford shopping center.  Recognize the last name?  



Are they comparing the removal of the robot from the Pet Shelter to the Boston Bombing, Orlando or Las Vegas??  Disgraceful!!!

Read this IPVM report for free.

This article is part of IPVM's 6,803 reports, 913 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports