Knightscope CEO on IPO, Opportunity & Controversy

Knightscope Prepares for IPO: A Deep Dive with Founder William Santana Li

After a considerable period of anticipation, Knightscope has officially announced a January 27th date for its initial public offering (IPO). The Mountain View-based company, established in 2013 by William Santana Li, previously a Ford executive, and Stacy Stephens, a former Texas police officer – also co-founders of the police vehicle company Carbon Motors – has achieved significant recognition nationally due to its advanced surveillance robots.

Current Deployment and Founding Principles

According to Li, approximately 80 to 90 Knightscope robots are presently deployed under contract across the United States, which currently represents the company’s exclusive market. Li, a native of New York, shared that the events of 9/11 profoundly influenced the company’s inception. He has also referenced the Sandy Hook shooting in prior discussions as a motivating factor preceding the firm’s establishment.

Steady Growth and Technology Adoption

Knightscope’s expansion has been consistent. Numerous law enforcement organizations have readily integrated the technology to enhance their operational capabilities. These robots operate continuously, verifying details such as license plates and even the MAC addresses of smartphones, aiming to both identify and prevent criminal activity.

Controversies and Public Perception

However, the company’s journey hasn’t been without challenges. Several publicized incidents, including a robot entering a swimming pool and another colliding with a young child, have garnered national attention. Li suggests that some of this publicity may have inadvertently benefited the company’s visibility. Organizations focused on electronic rights, such as the EFF, have voiced concerns, labeling the technology as a potential “privacy disaster.”

An Exclusive Interview with William Li

Ahead of the planned IPO, we had the opportunity to speak with Li regarding the company’s trajectory, future ambitions, and the controversies surrounding its technology.

The following interview has been edited for conciseness and clarity.

On Founding Knightscope and its Vision

TC: Could you share your background and the motivations behind founding Knightscope?

WL: I come from an automotive background, having spent considerable time at Ford Motor Company – a truly excellent learning environment. Over a decade, I held twelve distinct positions across four continents, encompassing all functional areas. I am immensely grateful for the experience. The impetus for starting Knightscope stems from both personal and professional factors. Building upon my automotive experience, I firmly believe that self-driving autonomous technology will revolutionize the industry. However, I diverge from the conventional path to commercialization.

Addressing Concerns About Accidents

TC: There’s an inherent expectation of accidents with real-world deployment of autonomous systems. The scale of potential impact differs significantly with multi-ton vehicles. Yet, several incidents have received widespread publicity.

WL: These incidents occurred several years ago, pre-COVID. Consider that a mature industry experiences approximately 15,000 accidents daily, having existed for a century. We experienced a few minor incidents with a novel technology half a decade ago; it’s important to maintain perspective.

The Inevitability of Incidents

TC: Is there an acceptance that accidents are likely to occur once these robots are publicly deployed?

WL: Absolutely. I’ve communicated this to the media, our underwriters, legal counsel, teams, clients, and investors: more incidents will happen. It’s reasonable to anticipate accidents. Evidence often suggests that human error plays a role. In many cases, the issue isn’t the robot itself, but rather unavoidable accidents. Will future incidents occur? Undoubtedly. My focus is on how we respond, conduct ourselves, support our clients, and ensure everyone’s safety, making necessary revisions.

Impact of Incidents on Company Profile

TC: Do these incidents significantly impact public perception of the robots?

WL: One could argue that. However, it’s also true that we’ve secured contracts from Hawaii to Texas to Rhode Island if these incidents were truly damaging. We’ve received renewals for our fourth and fifth years and have a proven track record of crime-fighting successes. Building a demonstration is relatively easy, and creating a prototype is achievable. But securing a client willing to pay full price for a year, renewing the contract, and adding more machines demonstrates genuine value.

Unexpected Benefits from Negative Publicity

TC: Do you believe the incidents ultimately raised the company’s profile and attracted more support?

WL: It’s not just a feeling; it actually happened. Even family members inquired if everything was alright. We did experience increased visibility and exposure.

Future Role of Robots in Law Enforcement

TC: Do you envision a future where the robots take a more active role in apprehending suspects?

WL: Absolutely not. The most effective approach is a combination of software, hardware, and human involvement. Technology should handle monotonous, computationally intensive tasks and provide deterrence. Humans are essential for law enforcement and decision-making. This hasn’t been achieved at scale before. To gain societal ‘trust’ in a new approach, avoiding offensive actions is crucial.

Robot Capabilities and Data Collection

TC: What specific information is the robot designed to detect?

WL: In the short term, they focus on key elements. This includes identifying individuals, particularly in commercial properties where no one should be present between 10 PM and 5 AM. Alerts can be sent via text, phone call, or user interface notification. We can process several hundred license plates per minute, enabling blacklisting or use for parking enforcement. Thermal scans are also possible. Stationary machines can perform facial recognition at ingress/egress points. Within a few hundred feet, they can detect mobile devices, treating your phone like a license plate.

Detecting Suspicious Activities

TC: Are there specific “suspicious activities” that trigger an alert?

WL: If someone has issued a BOLO for specific activities, those would be flagged. Looking ahead, I’m fascinated by the idea of figuratively giving a machine a dog’s nose – enhancing its ability to detect scents. In the near term, the primary technological hurdle is allocating time and resources to enable the machines to recognize sounds like breaking glass, car alarms, or gunshots, and then pinpoint their location.

Addressing AI Bias Concerns

TC: Are there concerns about AI profiling and potential biases?

WL: If you’re suggesting issues with racial bias – and I may be treading carefully here – I believe it’s a case of “garbage in, garbage out.” If you instill a belief that pistachio ice cream is bad during someone’s upbringing, they’ll likely continue to dislike it. Feeding an algorithm flawed or incomplete data leads to engineering input problems, not inherent technological bias. I’m optimistic that this will be corrected over time as engineers continuously improve their work.

The Importance of Human Oversight

TC: It’s important to acknowledge that implicit biases and profiling aren’t unique to machines. A company like yours must be acutely aware of this and work to mitigate it.

WL: Two points – part serious, part humorous. People worry about robots, but I worry about humans. The level of chaos across the country is alarming. Simple etiquette is also important. Before deploying robots, we advise clients to introduce them to the community – a robot lunch, a naming contest. When people understand the robot’s purpose, they’re more likely to accept it. If a client resists proper introductions, problems are likely to arise. Hollywood’s portrayal of robots over the past 40 years has undoubtedly influenced public perception. Communication, community engagement, and answering questions are crucial for building trust.

Distinguishing AI Bias from Robopocalypse Fears

TC: It’s important to differentiate concerns about AI biases from broader fears of a “robopocalypse.” While there may be some overlap, the concerns about implicit biases are valid.

WL: I’ve shared my perspective on that. I don’t have much more to add.

The Rationale Behind the IPO

TC: Why is an IPO the right move for Knightscope at this time?

WL: There are pros and cons to going public. We’ve been reporting to the Securities and Exchange Commission as if we were publicly traded for several years, but without the benefits. I needed to move the company “out of purgatory.” Having experience acquiring companies, I believe the public markets could provide opportunities for acquisitions. From a recruiting standpoint, having a publicly traded stock with a market value is more attractive than a phantom promise. Larger clients often prefer dealing with fully audited, public companies. And, of course, the company will require financing for continued growth. The global capital markets offer significantly more potential than what’s currently available in Silicon Valley.

Timing the IPO in a Volatile Market

TC: The market conditions aren’t ideal right now. Why proceed with the IPO at this moment?

WL: These processes take time. I’ll likely get in trouble for saying this… I’m not focused on short-term market fluctuations. This is a decades-long endeavor. Worrying about a temporary market downturn is unproductive. We have clients, a country we want to improve, and a need for resources to create jobs. We’ve proven our technology works and benefits society. Whether the market is up or down, we need to move forward.