Automotus Privacy Policy | last modified October 2023

 

Privacy is a growing concern for all of us, and our team takes privacy extremely seriously. Our technology has been private-by-design from the start. Unlike a surveillance system or traditional camera recording system, our optical sensors capture (1) de-identified sample image data to train our technology, (2) de-identified traffic and curb activity data to inform smarter policies, and (3) vehicle license plate information for the purpose of automatically processing parking payments and citations. License plate information is the only form of personally identifiable information (PII) we collect, and we never capture or share this information for any purpose other than to facilitate payment and enforcement of parking regulations, or to train machine learning models.

By design, our technology can only be used as a tool for cities to reduce congestion, emissions, and safety hazards as a result of the unprecedented rise in commercial vehicle traffic. It cannot be used as a resource for public safety authorities or police departments for any reason outside enforcing parking regulations, and the data we capture will never be sold to any third party. This policy outlines what our technology is (and is not), what information it collects, how we use (and do not use) it, and the steps we have taken both in the technology design and in our company practices to ensure we are protecting personal privacy in every possible way.

____________

​

Glossary
 

API: Application Programming Interface; Automotus’s API gives cities access to metadata to create more informed policies that help reduce congestion and emissions.
 

authorized user: Automotus employees and Automotus software development contractors who have each received training specific to managing PII.
 

de-identification: the process of removing any PII from an image that prevents a human or machine from extracting the identity of a person.
 

metadata: de-identified traffic and curb activity data that generates valuable statistics and insights cities can use to make more informed policies that help reduce congestion and emissions.
 

object data: inferences about a particular object observed by an Automotus camera system, including the object classification (e.g. pedestrian, bicyclist), location, and time of the observation.
 

personally identifiable information (PII): features of a person or vehicle that allow their identity to be extracted by a human or machine with a high level of confidence.
 

sample image: de-identified images collected by a sensor and transmitted to the cloud for training and validation.
 

web dashboard: Automotus's password-protected web application where customers can view aggregate object data and insights for their sensors.

 

 

____________
 

What is our technology?

​

Our team has developed first-of-its-kind computer vision technology that runs on cellular-enabled optical sensors secured to public infrastructure, such as streetlight poles. Each sensor contains a graphic processing unit and a cellular data chip. The sensor’s only requirement is uninterrupted power — all video processing is done locally and the processed data is then transmitted to the cloud via cellular connection.

Our technology is not a surveillance system nor a traditional camera recording system. Live video feeds are not remotely accessible to anyone other than authorized users at any time and access cannot and will not be granted to anyone else at any time.

 

____________
 

What data do we capture and how is it used?

​

Our technology captures three types of data to help cities manage congestion, emissions, and safety hazards caused by the unprecedented rise in commercial vehicle traffic.

We capture sample image data of vehicles and vehicle activity for the purpose of validating system performance and training the computer vision algorithms on our sensors to discern unique objects and activities, such as different vehicle types and double-parking. By design, our sample images are low in resolution and in quality, though we understand some PII may still be visible. As a result, our sample images always go through a process of image de-identification where all faces and license places are blurred beyond recognition before they are made available to any authorized user. Every authorized user has a secure login that is only issued after receiving specific training.
 

We capture de-identified traffic and curb activity data, often referred to as metadata, from video streams processed in real time and deleted instantly. This includes volume of parking and traffic; average dwell times; curb occupancy rates; and double-parked, over-time, and non-permitted parking violations. This data is then broken down by vehicle type - passenger, delivery van, freight truck, ride-hailing, bus, bike, e-scooter, etc. Our technology blurs all PII, such as faces and license plates, beyond recognition. We use this metadata to deliver aggregate statistics and insights to city transportation agencies who then use these to create policies that reduce congestion, emissions, and safety hazards in high-traffic areas.
 

We capture license plate information exclusively for the purposes of processing parking payments and citations, and training our computer vision models. We only capture license plate images of vehicles that use designated parking and loading areas monitored by our technology. We do NOT read or capture the license plate information of any vehicles that are not using designated parking and loading areas monitored by our technology. We de-identify all faces and non-relevant license plates in any images captured for payment and enforcement purposes, and all images are deleted immediately once the payment or citation is resolved.
 

We do NOT collect, transmit, or store any PII for any purpose other than facilitating payment and enforcement of parking regulations

 

____________
 

Can our technology be used as a resource for public safety authorities or police departments for purposes other than parking enforcement?
 

No, as a policy and by design our technology cannot be used as a resource for public safety authorities or police departments for any reason outside enforcing parking regulations. Our live video feeds are not remotely accessible to anyone other than authorized users at any time and access cannot and will not be granted to anyone at any time. 
 

____________

​

Do we sell our data to any third parties?
 

No, we never sell any data to any third party.

 

____________

 

How do we protect the information we collect from misuse?
 

Our technology encrypts all communications to and from sensors with TLS1.2 using industry standard AES256 encryption. Only authorized devices can communicate with our technology, and keys are carefully maintained and rotated frequently. This removes pathways for data interception or sensor access by unauthorized third parties. All of our data is stored in infrastructure hosted by Amazon Web Services (AWS) and access is restricted to authorized users using AWS best practices for identity and access management.

Additionally, our de-identification process has been audited and validated by third party auditors.

​

____________

​

What administrative safeguards do we have in place?
 

All individuals with access to vehicle, traffic, and parking/double-parking metadata undergo training and are provisioned access to this metadata using strict role-based access controls. Metadata never includes any PII. Access via API is restricted by access token, so only individuals who have an access token may access this data. Live video feeds are not remotely accessible, so access cannot and will not be granted to anyone at any time. Our internal security controls adhere to industry-standard best practices and have been audited and validated by third party auditors.

​

____________

 

How does our technology directly benefit residents?

​

Our technology directly benefits residents by providing cities with the insights and automated payment and enforcement solutions necessary to manage the unprecedented rise in commercial vehicle congestion and emissions. Armed with comprehensive, real-time traffic and curb activity data, policymakers are able to create smarter curbside policies that align with the real-time needs of all users, including pedestrians, cyclists, public transit users, and passenger vehicles, and minimize safety hazards and congestion at the curb caused by illegal commercial vehicle loading and unloading.
 

Additionally, by capturing vehicle license plate information, cities are able to equitably charge all curb users for parking and (un)loading. Traditionally, local residents are responsible for paying for parking, while commercial operators like Amazon, UPS and Postmates do not pay and instead park and (un)load illegally. With our technology, cities can now create designated (un)loading zones and automatically charge commercial vehicles for their time spent at the curb, reducing congestion and safety hazards.