Robots – Enhancing Our Security – Part 1

Some Questions I Ask:

• What needs to change for the US to become the safest country in the world? (02:27)
• How can you help achieve the security standards people expect? (08:10)
• How does the data collection system work? (17:02)
• What were some of the most surprising things that happened in the robot’s interaction with people that you didn’t expect? (20:32)
• Will your robots, at some point in the future, carry weapons? (25:45)

What You’ll Learn in this Episode:

• The extent of the shortage of police officers in the US (04:05)
• Knightscope’s robots capabilities (06:55)
• The price of Knightscope’s autonomous security robot services (14:40)
• How Knightscope’s robots communicate with people (19:36)
• The factors they took into consideration when designing their robots (27:20)

Connect with William Santana Li: 

• LinkedIn
• Knightscope
• Twitter

Connect with Ed Bernardon:

• LinkedIn
• Future Car: Driving a Lifestyle Revolution
• Motorsports is speeding the way to safer urban mobility
• Siemens Digital Industries Software

Ed Bernardon: The 2021 Global Peace Index ranked the United States 128th on the list of safest countries. Not a very good showing. Part of the reason may just be that Police departments and security companies are overworked and under-resourced. A need exists for more effective and far-reaching systems of security.

Ed Bernardon: One solution that’s gaining a lot of traction is security robots. The technology exists for autonomous robots to survey perimeters, collect pertinent data, and facilitate security operations. But can these robots actually help improve security in the United States and elsewhere? Well, Bill Li, CEO of Knightscope, has a company that makes autonomous security robots, and he has a goal of making America the safest nation in the world. Now, these aren’t the diabolical, armed Robocops that we see in the movies, but rather, you could think of them as assistants to your local police and security companies. Bill is going to tell us a little bit about how a former Ford executive ended up in the Robocop business and all the technology that’s behind his autonomous machines.

Ed Bernardon: My guest today, Bill Santana Li, is the Chairman and CEO of Knightscope, an advanced security technology company that makes fully autonomous security robots, built to deter, detect, and report. In Part 1 of this two-part interview, Bill and I discuss the security needs of America in this day and age and how Knightscope security robots can help meet those needs. We talk about the functionality of Knightscope’s robots, data collection, and the inspiration behind Knightscope, and everything in between. Join me, Ed Bernardon, on this episode of The Future Car as Bill and I take a look into the future of robotic policing.

Ed Bernardon: Bill, welcome to the Future Car podcast.

William Santana Li: Thanks for having us, and greetings from Silicon Valley.

Ed Bernardon: Knightscope. So, it’s an advanced security technology company; you make robots to deter, detect, report crimes, and your mission is to make America the safest country in the world. So, from your perspective, what needs to change in the USA for it to become the safest country in the world?

William Santana Li: It’s actually rather simple. It’s to provide the 1 million officers and 1 million security guards with actual tools for them to do their jobs much more effectively. We would never dare treat a soldier in a theater of war in the manner that we treat our officers and guards. The two-plus million soldiers out in a theater of war get every level of technology you might ever imagine or can’t possibly imagine. And there are massive companies like Lockheed Martin, General Dynamics, Raytheon, or Boeing to build a soldier whatever he or she might need on the battlefield or in a theater of war. And there’s an actual process: there’s risk capital; there’s someone in charge, a secretary of defense; there’s a whole military complex that builds all this stuff — it takes too long, it costs too much money, but it actually delivers the widget, be it a new submarine, new jet fighter, new tank, whatever you might need. That doesn’t actually happen on our own soil, so the US Department of Justice and the US Department of Homeland Security have no federal jurisdiction over the 19,000 law enforcement agencies and 8,000 private security firms. There is no process, there’s no risk capital, there’s no innovation strategy, there’s no accountability, and that’s why in the year 2022, you’ve got security guards out there literally in the parking lot with a number two pencil and a notepad. And that has got to change.

Ed Bernardon: So, unlike the military, where there’s a centralized way to get funding and look at all the needs, you’ve got 19,000 police departments all over the country that are taking care of themselves. They can’t go out and do a big program to advance the next high-tech widget to help them?

William Santana Li: No, they’re understaffed and under-resourced to do the job that’s at hand in the first place. So, them trying to come up with a next-generation technology or strategy is not within their remit or resources. Maybe it’s not a popular thing to say, but the country is understaffed and under-policed. So, I mentioned 2 million folks, they’re running 24/7. So, at any given time, you can’t triple-shift a human. If you want to cover 24/7, you need four people, so you need to divide the 2 million by four. So, at any given time, there are only 500,000 humans trying to secure 332 million Americans across 50 states. That basic, simple math doesn’t work, and I can prove it doesn’t work. If you just go to the FBI crime clock, you see a horrendous amount of violence and property crimes every few seconds in our nation; therein lies the evidence of what we’re doing is not working, and I think it’s time we fix it.

Ed Bernardon: You probably have done this, but if you were to survey a group of police officers and you were to say to them, “I can create a technology for you to make your job easier,” what do you think they would say? Or what do they say?

William Santana Li: Well, anything that’s going to keep an officer out of harm’s way, one, is really important; I think, two, giving them unprecedented situational awareness to understand what’s going on in multiple locations at the same time; I think, third, it’s a little soft side of things, but something to disintermediate or deescalate a situation, because it’s not always the best idea — and the officers who would agree with this — to show up uniformed officers that are armed to a particular situation. It might actually be more helpful to have a step before that. And also the officers don’t have enough time and can’t be everywhere at the same time. So, wherever they can be deployed to be most effective would be much more helpful. And what they really need is some tools. It’s not as much as everyone wants to talk about Robocop and Terminator, and all this other silliness, and “the robots are coming to kill everyone and take everyone’s job.” I mean, this is a serious time in the country’s history and we need serious solutions, and we’ve developed something that can actually have a big impact. And I think having my co-founder actually be a former officer, and we have multiple military former law enforcement on the team, the folks that you speak of are actually on the Knightscope team and help dream this up.

Ed Bernardon: I ask a question sometimes to companies and people that work in places that make autonomous cars: “Is your autonomous driver ever going to be as good as a human driver?” And one of the things they say is, “Well, eventually they will. And one of the reasons is, there are so many sensors looking at all different directions that it’s like a human driver with seven heads that can see in the dark.” And in a way, this seems like this big benefit you’re talking about is giving your police officers or security team more eyes and ears so they can see what’s going on in places where they might not be readily able to see.

William Santana Li: Eyes, ears, and voice. We have 360-degree eye level — so we’re not staring at the top of someone’s head — 360-degree eye-level high-definition video, both live streaming and recorded. The machines can read multiple 100 license plates a minute, can do facial recognition, can run a thermal scan of an environment, can detect a mobile device in an area. But the officers can also speak through the machines as if it’s a mobile PA system, or have a two-way conversation human to human. So, again, it’s eyes, ears, and voice on the ground in multiple locations at the same time.

Ed Bernardon: So, the security business has been around for a long time. And I’m sure there are certain standards that people expect from their security patrols or security systems. So, tell us a little bit about how your robots improve this. First of all, how can you help even achieve the standards they expect and maybe even go beyond that? And maybe tell us a little bit about each of the robots and how they fit into this.

William Santana Li: The standards that the nation expects, and I believe every citizen expects, is I don’t believe the founders of the country ever expected us to build a society where going to work, going to school, or going to a movie theater literally came with a risk of being shot or killed. So, that’s kind of a basic, very low standard that we need to be able to fulfill. And as a country, we’re failing miserably. I think it’s six months into the year, almost 350 mass shootings. That’s just not acceptable, untenable, and we need to fix it. By the way, who’s gotten fired over all those? No one. Why? Because no one’s accountable. That’s another big issue here. To answer your question, I think one of the things that most people think about is what we just discussed, is the eyes, ears, and voice on the ground; they can get a better understanding. But the less maybe glamorous part is just physical deterrence. And I’ll explain this to your audience in a couple of ways. Just very simply, if I put a marked law enforcement vehicle in front of your home or your office, criminal behavior will change. It’s no different than you driving down the highway and you see a marked law enforcement vehicle on the side of the road, frankly don’t care what speed you’re doing, you are going to pump those brakes, look down at the speedometer, or whatever it is, but you will have a reaction. So, if you’re trying to steal a car in a parking lot at three o’clock in the morning, and you pull up and there’s a five-foot-tall, 400-pound machine roaming around — there’s no one remote controlling it, this is fully autonomous — this robot is going, it says “security” or “police” on the side, it’s making some sounds, you have no idea what it does — you’re going to go steal the car somewhere else or not steal the car at all. And that’s literally what’s been happening with our clients, is just physically being there stops a lot of the negative behavior in the first place.

Ed Bernardon: Yeah, it’s knowing that there are some eyes and ears on you. You raised a good point, though, about accountability. I never thought of that. If no one’s accountable, then what’s the motivation to improve?

William Santana Li: Citizens are not willing to accept the status quo anymore. We just took the company public earlier in January, just before the public listing, with 35,000 investors participating as a part of Knightscope’s growing story over the last nine years. You don’t get that outpouring of 35,000 people going “I had enough of this, we need to fix the problem.” We still live in a democracy and we need to hold the institutions accountable for what they’re responsible for. What is the first role of government? To protect its citizens. The country is well over 200 years old, we’re in our 46th president, and no one’s actually fixed this problem. And the problem being, to be absolutely clear, crime and terrorism have a $2 trillion negative economic impact on the US every single year. It’s a hidden tax we all pay in blood, tears, and treasure. And I think this is an opportunity for technology and the private sector and public sector to partner together to actually fix the problem. And if anyone’s doubting that a combination of autonomous technology, robotics, AI, and electric vehicles can actually solve the problem, I’d just submit to you the results. Go to knightscope.com/crime, and we’re just getting started, you can see a nice long list of all the positive impacts the technology has already made on society, and we’re just getting going.

Ed Bernardon: Because of what’s happening in society, have you seen the private sector being more motivated now to try and improve? And has it helped you in showing them how you can help them?

William Santana Li: Oh, absolutely, by far. As I mentioned, my co-founder is a former law enforcement officer. We actually started selling the technology into the private sector first. In my opinion, it’s ill-advised for young technology companies to do business with government sales as the first go-to-market; you will run out of money; you will fail before you get your first order. And unfortunately, both law enforcement and security industries are not the most technologically savvy sectors or folks headfirst. So, we actually did start on the private sector first — so, primarily around hospitals, commercial real estate, corporate campuses, casinos, manufacturing plants, logistic facilities, basically anywhere indoors or outdoors, mostly outdoors, where you might see an officer or guard is an opportunity for Knightscope. So, some of our clients are Bank of Hawaii, Citizens Bank, Houston Methodist Hospital, Dignity Health, numerous casinos, Samsung — there’s a whole host of folks in the private sector that have shown that this can work. And then now we’ve started adding more law enforcement agencies like the Huntington Park Police Department that has done very well with the technology and continue to renew the contracts. So, we’re excited about the future.

Ed Bernardon: Some of those places you mentioned, like casinos and hospitals, they’re not really the types of places that are used to having robot technology.

William Santana Li: One little characteristic before we had gotten into it because I never thought that that would be a place where we would spend a lot of time. If you sit back and just think about it, both the casino and hospital, 24/7 stuff is going on. In one, there’s alcohol and cash involved, so drama shall ensue. The other one, it’s a hospital, so there’s definitely a lot of activity there 24/7. And obviously, during a pandemic, you likely don’t want to be the security guard at the hospital. So, there are a lot of opportunities. I think, if you look over the next three decades of what we intend to do, criminals and terrorists don’t care that we work well in casinos or hospitals, or corporate campuses. Basically, they can be anywhere, so we need to be everywhere. So you might imagine that we would have every size from ultra small incognito mode at a federal courthouse through the largest size to get through a city and highway so that we can secure all the places you visit, live, study, and work.

Ed Bernardon: Now, you provide your robots as a service rather than actually selling the robots. So, if I was running a hospital, I don’t have to worry about maintaining my robots or programming them or doing any of that, that’s all handled by you. So, that makes it a lot easier.

William Santana Li: All-inclusive service. We price it so that it’s very difficult to say no. At $3 to $9 an hour, it can be very enticing for those that are budget constrained. We can run single, double, or triple shifts. We take care of everything: we build the machines, we deploy them, we support them, we service them 24/7, 365. Anything goes wrong, it’s one throat to choke, we’re responsible. Because our clients are busy, they have their operations they need to be running, and we want to make this as affordable and cost-effective as possible. And I think it’s proven itself well; numerous clients that have renewed for the third year, the fourth year, the fifth year because we’ve actually had a positive impact.

Ed Bernardon: You’ve got several different models: K1, K3, K5, working on the K7. Tell us a little bit about what each one of them does and how their applications differ from one model to the next.

William Santana Li: The K5 is the most popular model, that’s what I was alluding to earlier. It’s primarily used outdoors, again, five foot tall, three foot wide, 400 pounds, roams around autonomously 24/7, including autonomously recharging itself.

Ed Bernardon: Is that like a Roomba?

William Santana Li: It’s more like a self-driving car, where we use LIDAR, sonar, and a bunch of crazy software to be able to autonomously operate in dynamic, indoor and outdoor environments. So, think pre-COVID, indoors, at a mall at three o’clock in the afternoon on a Saturday is complete mayhem; you’ve got to be able to manage to get through those crowds, or several locations where you might be at a nine-story parking structure going up and down with plenty of vehicles and plenty of humans. So, that’s our most popular model. There is a K3, which is a little bit smaller for indoor applications: four feet tall, two feet wide. And then the K1 is a stationary version, so we can cover the same technology stack for ingress/egress location — so, wherever a lot of humans might be coming in and out or a lot of vehicles — think of a casino where you could put that at the entryway or exit to making sure that you’ve got enough coverage to better secure the facility.

Ed Bernardon: And you also have a system for collecting data because that’s one of the big things that these robots do as part of surveillance with all their different sensors. So, tell us a little bit about that system.

William Santana Li: Sure. That’s how we get Big Brother to work, is collect as much data as possible so we can surveil everyone and sell all the data to the highest bidder — just kidding. A lot of folks worry about Big Brother and “What are you doing with the data?” Contractually, the way we set it up is real simple, basically, the security data is owned by the client, not by us. So, we are not in the position to be selling data to the highest bidder or monetizing it, or doing something else with it. Our mission is to secure the country, not to go monetize someone’s data. So, that’s the first point. The machines do generate a massive amount of data; it’s 90 terabytes per machine per annum that no human is going to be able to ever process. So, we put that in a digestible format in a browser-based user interface we call the KSOC. As everyone focuses on the robots, the magic is actually on the other side, which is the Knightscope Security Operations Center; this is where our clients typically utilize the technology to interface with the machines, grab historical data, pull as much live data as they want. If something’s happening at the moment, then that gives them the eyes, ears, and voice on the ground that we spoke of earlier.

Ed Bernardon: And the confidentiality of data is very, very important. So, it sounds like you’ve got that covered.

William Santana Li: Yeah, we’re operating in public environments. We’re not in someone’s restroom or conference room or office or what have you, so that’s not where we operate — one. Two, if we didn’t have the basics covered on privacy, we wouldn’t be able to operate on a Fortune 500 corporate campus, a financial institution, a hospital, or a law enforcement agency. I think, third, we’re about 17-18 months into a two-year little bit of a nightmare US cybersecurity review process with the US federal government. And hopefully, by the end of the year, we’ll get an ATO or an Authority To Operate. So, if we’ve gone through, at that time, two years of scrutiny by the US Federal Government and now we can work with our sponsor there, which is the US Department of Veterans Affairs, I think that should give some comfort to folks that we’ve thought through the processes and the technology carefully.

Ed Bernardon: Now, you said these robots can actually talk to people as well. I imagine it’s not the actual robot that’s talking but a human talking through the robot. Is that how that works?

William Santana Li: Both. So, I’ll give you a couple of examples. Let’s say you’re a commercial property owner. Between 10 o’clock at night and five in the morning, there should be no one waltzing around your property. If we see someone, we could have the machine do an automatic pre-recorded broadcast; “Hey, it’s 02:00 in the morning, you’re trespassing, I’m calling the authorities.” You could have some fun with it: You’re at the mall, you’re driving by Starbucks, “Hey, you want a coffee?” It could be Happy Mother’s Day, it could be random, it could be the time of day, or it could be a human on the other side either trying to greet someone or deescalate a situation — like literally a talk down, which is, “Hey, it’s 02:22 in the morning. We’ve captured your information, and if you don’t vacate the premises immediately.” And then the third one is text-to-speech; the officer or guard can sit there and type what they want, and then the machine can announce that.

Ed Bernardon: So, they interact with people. Whenever you interact with people with a machine, you never know what’s going to happen. I want you to turn the clock back to the early days when you first started putting your robots out there, what were some of the most surprising things that happened in this interaction with people that you didn’t expect?

William Santana Li: For the Star Wars fans out there. The company was founded in April of 2013, we put the first one out in the wild on May the Fourth of 2015 — so, may the force be with you. And honestly, the team, we were very concerned internally because no one’s actually done this in the real world. We didn’t know if society would allow us to do this — like what’s going to happen? We’re going to go put this machine out there. Are people going to go nonlinear or it’s going to be a non-event or whatever? What I didn’t expect to happen was we put the machine out there and it just drove so many robot selfies; people just wanting to take a picture with the robot; parents driving their kids for four hours just to take a picture with the robot; girls leaving lipstick, literally, marks on the machine with a big kiss on the side. I didn’t expect that, to be honest, at all, that was the first couple of things. The other one was someone had asked me, “Hey, are you going to name the robots or brand the robots?” And I’m like, “It’ll be fine, we’ll just put Knightscope on it, leave him that color and it’ll be fine.” I want to say 90% of our clients have the machines branded with their own brand on it — like it literally says “Samsung” on the side, which I didn’t expect. And then almost all of them are named, meaning the client had a robot naming contest. And then there’s a different emotional attachment with the people there and the robot because now they named it, and now it has a personality, for them. And then there’s a different bond that gets built, as opposed to “the scary robot is coming here to kill everybody.” That’s a couple of things that I really didn’t expect.

Ed Bernardon: And you certainly didn’t expect that lipstick would make it so your security camera couldn’t see anyone after one of your robots gets a kiss. So, these are the things you learn when you start doing testing in the real world. You said that people name the robots, and I heard there was one named Rosie that I think fell into a river. What are some of the mishaps? Tell us a little bit about Rosie.

William Santana Li: First and foremost, as a former auto executive, you would agree with me that the automotive industry is a mature industry; it’s 100-plus years old. Pre-COVID, how many accidents occurred every day? Literally 15,000 accidents every single day for a 100-year-old mature industry. We’ve had a few incidents over the last few years and there will be more. The issue here is that you cannot build this technology in some laboratory somewhere. No one’s ever done this before, so you have to have that field experience. And now that we’ve been through a million and a half hours of actual field experience with paying clients 24/7, in the rain, at night, across the entire country, through six summers and six winters; we kind of learned a thing or two because we’ve seen a thing or two. And I expect that’s why insurance exists as an industry, accidents occur. In some cases, it’s our fault, and the cool thing is we drop a new software every couple of weeks, new hardware every few months, and we learn over time and the technology gets better and better. But at some times — and now we have the evidence to back this up — it’s not the technology’s fault or the robot’s fault, it’s the human’s fault. A lot of folks shouldn’t have driver’s licenses, let’s just put it that way. If it was you that had nothing to do with being a robot or not, you’d be run over. People get in their car, don’t look in the side view mirror, the rearview mirror, just put it into reverse and floor it. There’s no level of technology we can build — I can’t fix stupid. And that happens a lot more often than I was anticipating, to be frank. It’s funny a lot of folks in the media and society are worried about robots. From where I sit, given what I just said to you and all the crime going on, I don’t worry about the robots, I frankly worry about the humans. The humans don’t create all this violence and mayhem and accidents; it’s certainly not the technology.

Ed Bernardon: I mean, a robot that acts on its own in a factory is skilled — if we can use that word — differently than a robot that has to interact with a human while it’s trying to do a certain task because humans are unpredictable, so it’s not easy. Some of the jobs I’ve had in the past were more towards development and R&D; one of my engineering bosses told me once, “Ed, it’s okay to make mistakes as long as they’re new mistakes.” So, you don’t want to make the same mistake twice. And like you said, you can’t anticipate everything, so now you have a windshield wiper to wipe the lipstick off

William Santana Li: The stuff that we all worried about, in a lot of cases, didn’t happen. And the stuff that wasn’t ever on a list, was never in a meeting, never one ever worried about is the stuff that ends up causing all kinds of ridiculous issues. That’s why that field experience is really important.

Ed Bernardon: Now, your robots don’t carry weapons. And I’m sure you probably get this question all the time. Is this something maybe you’re planning for the future? Could that be an evolution?

William Santana Li: Absolutely not for at least three reasons. The first reason is that subject is covered already by the Department of Defence. As I said, the DOD has over an $800 billion budget; why would you want to be investing more in that space? It’s not what we need to be doing to secure our country. Second, if you’re trying to build a new technology for society to trust, the last thing you need to be doing is tazing someone. And then third, half in jest and half in serious, is I’ve got no engineers interested in being the test dummy to go develop that stuff. So, for us, bright red line, absolutely not. Our stated goal is, over the long-term, we want to put a million machines in the field to help a million officers and a million guards actually be able to do their jobs appropriately and effectively.

Ed Bernardon: You mentioned military machines, there may be military machines that are autonomous that could carry weapons. Are there any commercial machines out there similar to yours, commercial surveillance robots that do carry weapons?

William Santana Li: Not that I am aware of. There’s always stuff on YouTube or in someone’s garage.

Ed Bernardon: Yeah, but not commercially available.

William Santana Li: Not that are autonomous. There’s a lot of tactical stuff that’s used remote-controlled, that can be used for that purpose, but not at scale, not commercially viable that I know of.

Ed Bernardon: Plus your robots are so friendly looking. It looks like an R2-D2 with more of a pointy head. That’s why people want to give it a hug.

William Santana Li: Yeah, so we had a little bit of a design challenge here because this is not on a battlefield; you need to operate in society. You need to make it approachable enough that you’re not going to scare grandma or the child. But you need to make it have enough of a commanding presence or a minor level of intimidation that it’s actually going to help with stopping the negative behavior. We had to think through every surface, every font, every color, that we use, every radius, and every little thing on the machine to give you that notion. But at the same time, there’s an audio component that goes with it, there’s a patrol sound that you don’t want to sneak up on someone but you also want to deter that behavior by, like, “I don’t know what that sound is, let me go somewhere else.” It’s a little bit of social experimenting and design at the same time.

Ed Bernardon: Everybody loves R2-D2. Have you ever thought of teaching your robots to speak whatever language it is that R2-D2 speaks? That would actually be a friendly interaction don’t you think?

William Santana Li: The machines can speak in different languages, especially if it’s a pre-recorded type of thing. So, we do have some clients that I believe most of it has been in Spanish, in particular locations where they would like a dual language machine, and also picking the gender is important for some clients. So, that’s also part of the onboarding process.

Ed Bernardon: Tell us about your favorite Knightscope crime-fighting story that your robots have been able to do.

William Santana Li: So, a little preamble before I answer the question. For most officers and guards, 95% 98% of time, there’s nothing going on. It’s not like what’s on TV all day long. It appears as though it’s chaos and drama every moment of the day, that’s not the reality. So, the statistical probability of us being somewhere at the right time at the right location and we are able to help is close to zero — the likelihood that something happens. And that, to me, is most fascinating that we’ve been able to have any crimefighting to speak of in the first place. I think the one that was emotional for me was, I think I was in a car in New York City going from one meeting to another, one of my colleagues calls me up, he’s like, “Hey Bill, you’re not going to believe what happened.” This was, I think, our first win. He said, “We just helped the police department issue an arrest warrant for a sexual predator.” I nearly dropped the phone. I can’t elaborate too much, but eyes, ears, and voice on the ground to give the officers and guards an ability for them to better understand the environment where they might not always be. So, we had all the evidence to prosecute. And to me, having a crimefighting win was awesome. But also for years and years, you have no idea how hard it’s been to build this company; “Hey, Bill, you’re out of your mind, this will never work. It’s hardware, it’s software, it’s too complicated. Physical security is not an investment thesis. This will never work. And then this will never work, and this will never work.” And then to sit there and go, “Here you go. It worked.” And then to have it happen again and again, that’s the one that probably sticks out in my mind, that first one. I was literally speechless, I’m like, “I can’t believe we just did that.” I get most excited about that. Technology is really cool, but to be able to actually have a positive impact is invigorating.

Ed Bernardon: That’s part 1 with Bill Santana Li of Knightscope. Join us on our next episode when we’ll continue our discussion on how real Robocops will help make a safer world. And as always, for more information about Siemens Digital Industries Software, make sure to visit us at plm.automation.siemens.com. And until next time, I’m Ed Bernardon, and this has been The Future Car podcast.