robots help neighborhood crime watch

The Rise of Robo-Watchdogs

The concept of using robots for neighborhood surveillance first gained traction in the mid-2010s. As advances in robotics, computer vision, and artificial intelligence accelerated, it became feasible to build autonomous security robots capable of navigating city streets and detecting anomalous behavior.

Some of the early pioneers in this space included:

• Knightscope: Founded in 2013, this Silicon Valley startup began deploying its iconic “K5” security robot in 2015. Resembling a squat, white R2-D2, the K5 patrols malls, parking lots and corporate campuses.

• Gamma 2 Robotics: This Colorado-based company introduced its “Ramsee” security robot in 2016. Ramsee uses lidar scanners and thermal cameras to patrol indoor spaces like warehouses and data centers.

• Robotic Assistance Devices: RAD launched its “SCOT” (Security Control Observation Tower) robot in 2018 for use in airports, prisons, and energy facilities. The SCOT is a stationary tower equipped with cameras, speakers, and other sensors.

As these commercial security robots proved their worth, police departments and neighborhood watch groups began to take notice. The idea of a tireless, ever-vigilant robot keeping an eye on the streets 24/7 was appealing as a force multiplier to supplement human officers and concerned citizens.

How Robo-Watchdogs Work

While designs vary, most neighborhood watchdog robots share some common features and capabilities:

• Autonomous navigation: Using GPS, lidar, cameras, and internal mapping, the robots can patrol a pre-defined area along an optimized route without human piloting.

• 360° sensing: An array of high-resolution cameras, thermal imagers, and microphones give the robot a full view of its surroundings in any lighting conditions.

• Anomaly detection: Onboard AI algorithms analyze the video, sound, and other sensor data to identify suspicious objects, activities, or people based on programmed parameters.

• Deterrent presence: Standing 4-6 feet tall and emblazoned with police decals, the robots’ physicality can discourage crime and loitering. Strobe lights and sirens can be activated as an additional deterrent.

• 2-way communication: Loudspeakers and cameras allow an operator to remotely talk down a suspect or communicate with a crime victim before first responders arrive.

• Evidence gathering: All video and audio is recorded and securely streamed to the cloud, providing an evidence trail and reducing the need for eyewitness testimony.

When the robot detects a likely crime in progress, it alerts its human supervisors back at the police station or neighborhood watch command center. The human can then review the evidence, communicate with people at the scene if needed, and dispatch police officers to the robot’s GPS location.

Robo-Watchdog Deployments

One of the first police departments in the US to deploy a purpose-built neighborhood watchdog robot was the Huntington Park Police Department in Los Angeles County. In June 2019, they began a pilot program using a “HP RoboCop” robot to patrol Salt Lake Park.

The RoboCop, which cost around $70,000 and was funded by private donations, had an immediate impact. “It’s working extremely well so far,” said Huntington Park police chief Cosme Lozano in an interview. “In the first month, we saw a 28% reduction in crimes reported, a 27% increase in arrests made, and a 38% increase in citizen-reported suspicious activity.”

Bolstered by this success, other cities soon followed suit with their own robo-watchdog deployments:

Even in major cities and high-crime areas, the robo-watchdogs were proving effective at deterring criminals and serving as a force multiplier for police.

Controversy and Criticisms

Not everyone welcomed the arrival of robotic police assistants on city streets. Some citizens and civil liberties groups raised concerns about privacy, bias, and the potential for abuse.

The ACLU argued that the always-on cameras and microphones of robo-watchdogs amounted to indiscriminate surveillance of the public. “These robots may deter some crime, but they come at the cost of constantly monitoring innocent citizens going about their lives,” said ACLU attorney Chad Marlow.

Others worried that the AI software could reflect racial and socioeconomic biases in what it deemed “suspicious.” MIT computer scientist Leah McMillan said, “We have to be very careful that we’re not just automating and reinforcing existing police biases with these robot watchdogs. The neural networks need to be rigorously tested for fairness.”

Some citizens felt the robots were an Orwellian overreach. At a Huntington Park city council meeting, resident Jenna Perez said, “I don’t want a creepy robot rolling up and down my block all night peering in windows. It feels like we’re living in a police state.”

Police departments listened to these concerns and took steps to address them. Chief Lozano emphasized that the robo-watchdog video is only viewed by human officers and not routinely monitored or stored long-term. Police also invited community members to a demonstration showing exactly how the robot and its AI work to dispel misconceptions.

“Our robot isn’t going to stop or interrogate anyone,” assured Chief Lozano. “It’s not doing facial recognition or accessing databases. It’s there as an extra set of eyes to help keep the neighborhood safe. If it spots trouble, it alerts a human officer who decides what action to take.”

The Future of Robo-Policing

As more cities experiment with robo-watchdogs, a number of trends and open questions are emerging that will shape the future of the technology:

• Specialized robots: Purpose-built security robots like Knightscope’s K5 currently dominate the market, but some experts predict a shift toward smaller, cheaper, more specialized bots. Swarms of drones and insectoid robots may one day assist police as mobile cameras and sensors.

• Expanded automation: Today’s robo-watchdogs are largely remote-controlled by humans, but continuing advances in AI could allow future models to navigate, assess threats, and even interact with citizens more autonomously. The role of the human operator may evolve more toward high-level mission planning.

• Addressing bias concerns: As facial recognition and predictive AI improve, they will likely be deployed more widely in police robots, which will intensify debates around algorithmic bias. Extensive bias testing, transparency, and public input into training data and model design could help build trust.

• 5G and edge AI: High-speed 5G networks will allow robo-watchdogs to stream higher-quality video and process more data using edge computing. Onboard AI chips from Nvidia and others will give the robots more intelligence to detect anomalies in real-time.

• Changing beat cop role: As robots take over routine patrols and surveillance, human officers may shift toward more high-level crime-fighting tasks that require human judgment and social interaction. Robots could be a tool to let cops focus more on investigations and community relations.

The long-term vision for many in law enforcement is a “robot-assisted police force,” with AI watchdogs serving as tireless partners to human cops on the beat. Such a force could allow police to cover more ground, respond faster to crimes in progress, and gather more evidence, all without putting more officers in harm’s way.

Whether that vision comes to pass will depend on how well the technology addresses legitimate citizen concerns around privacy, fairness, and responsible use. Police robot makers and departments have more work to do demonstrating that robo-watchdogs will remain firmly under human control and judgement.

“We can’t expect robots to solve all our crime problems or replace human police,” said Lozano. “But if we do this right, with proper policies and safeguards, these robot assistants could be a valuable tool to amplify our neighborhood watch efforts and make our communities safer.”

Frequently Asked Questions

1. Are the robot watchdogs armed?

No, the neighborhood watch robots are not equipped with any weapons. They are unarmed and cannot use force. Their role is to observe, report, and deter crime with their presence, not to intervene directly.

2. What data do the robots collect and is it kept private?

The robo-watchdogs have cameras, microphones, and other environmental sensors to monitor their surroundings. Video and audio are live-streamed to human police operators but are not routinely recorded or stored long-term unless needed as evidence. Personal data like faces are blurred in the feed. All data collection adheres to the department’s privacy policy.

3. Can the robots stop or interrogate people?

No, the robots are not permitted to initiate interactions with citizens or stop/interrogate people. If the robot detects a likely crime in progress, it alerts a human officer who then responds according to police protocols. The robot itself will not confront or apprehend anyone.

4. How much do these police robots cost?

Costs vary, but a typical security robot like the Knightscope K5 has a sticker price around $70,000. However, most robots are leased on an annual contract that includes maintenance and software for between $4,000-$8,000 per month. Cities often fund robot acquisitions through a combination of government budgets, asset forfeiture funds, and private donations.

5. What prevents the AI from being biased in who it deems ‘suspicious’?

This is an important issue that robot makers and police are taking seriously. The AI algorithms are trained on diverse datasets and rigorously tested for bias before deployment. Citizens are involved in the training process to provide input on what behaviors are truly suspicious. The AI also focuses more on actions than personal attributes to avoid improper profiling. Suspicious predictions are always reviewed and confirmed by a human before any action is taken.