In May, San Francisco became the first U.S. city to ban the government use of facial recognition technology. Shortly after, Somerville, Massachusetts became the second U.S. city to pass a ban on face recognition surveillance technology, and last week Oakland became the third U.S. city to pass a similar ban. Many U.S. cities will be voting on similar bans in the coming months, in an attempt to halt the erosion of privacy and preserve democratic freedoms. Both citizens and local governments have increasing concerns about the application of this new technology.
What is facial recognition technology?
Facial recognition is a technology that uses computer algorithms to look for a general face pattern. I want to make a very clear distinction between facial recognition technology, which detects the presence of a face, and systems that match specific facial patterns in order to identify specific people. Much of what is described as “facial recognition technology” is actually facial recognition within the context of searching for a face in a database. I’ll refer to this as “facial identification”. Facial identification goes a step further than facial recognition in that it creates a “face token” to represent a face by identifying the distinct features of the face, including: a rectangular region that contains a face, the length of jawline, the distance between the eyes, and nose width, in order to distinguish one face from another. This face token can then be compared to previously generated face tokens in a large database, in order to identify specific individuals.
Facebook holds the largest facial database. As millions of users upload photos to the social media site each day, Facebook’s facial recognition software, DeepFace, scans the photos for faces and recommends tags. Facebook is currently only using this software to tag individuals, but that may change. The data is valuable, and Facebook’s terms and conditions may change to allow for other uses, including government or third party marketing activities. The concern is that both government agencies and private companies would have access to an individual’s face token, and the individual would lose control of their personal identifying information. It is important to note that the recent bans on facial recognition software described above restrict use by police and government, but they don't stop private companies from using this technology.
I believe that the concern about the use of facial recognition technology for most people is identification for the purpose of surveillance and tracking. This makes many of us uncomfortable, and evokes feelings of big brother and a mass-surveillance state.
In my opinion, a ban on facial recognition technology is not the answer. It is possible to utilize this important technology, while also taking measures to ensure its responsible application, including regulation. In fact, facial recognition technology can be used to protect privacy. Any regulation should be driven by a non-partisan partnership of government, citizens, and private companies, with thoughtful and ethical guiding principles.
Facial recognition for good
The act of identifying an individual using the pattern of their face is valuable in specific use cases, including: identification documents such as driver’s licenses and passports, increasing public safety and security, and identifying lost children. In New Delhi, for instance, facial recognition systems helped trace almost 3,000 missing children in 4 days. Without question, it is important for government agencies to have the ability to quickly confirm an individual’s identity in specific situations.
At Radical I/O, we only engage in ethical, transparent technology development. We are a different kind of technology company - one that values community, relationships, and social good. We work with government, ethical enterprises, and education organizations to make technology and data accessible and useful to all, while safeguarding personal privacy.
De-identification to protect privacy
We recently created a software application that uses facial recognition technology to protect personal privacy. Our application detects and blurs personal information from videos and images, including faces, license plates, and other identifiers. When a video or image is collected, a machine learning algorithm recognizes the identifying data, and it is immediately blurred, and a new video or image is created in its place. This allows the video or image to be free of personally-identifiable information.
Cities and government currently gather and use data from multiple sources - street sensors, beacons, traffic patterns, communications, and even directly from their citizens. This information provides insights that are used to inform decisions on city infrastructure. These data-driven decisions create more livable and sustainable communities. However, these efforts can be impeded by the very real concerns cities have about adhering to privacy laws, as well as growing citizen concerns. Cities and government can, and should, take the extra step to ensure the privacy, consent, and data security of its citizens, including by adopting de-identification technologies.
The de-identification software that we have created at Radical I/O could be used in a variety of ways. For instance, when citizens report issues or request a service on a government website, they may include a video or photo along with their issue or request. As the image or video is acquired, any personally identifiable information, which may have inadvertently been included, would be immediately blurred. The submitted issue and accompanying media can then be used for review and action on specific issues and infrastructure decisions, including traffic pattern analysis, parking management, and smart crosswalks. If we can ensure that citizen data is protected, cities can feel confident in adopting new technologies in order to improve city services and operations.
Let’s communicate, collaborate, and regulate
The discussion around privacy, consent, and security is an important one. I believe that it is ultimately the government's responsibility to protect their citizens, and this should include their right to personal privacy, data security, and limitations on the use of their data beyond clearly defined purposes. We need to protect democratic freedoms and human rights, ensure against the potential for abuse, and mitigate other societal risks. This requires ethical, transparent laws and regulations that have been established, adopted, and will be continually evaluated by citizens, government, and industry.
I would love to continue this conversation. Please feel free to contact me if you would like to discuss responsible application of facial recognition technology, provide feedback, or if you would like to demo our new de-identification functionality on our SimpliCity™ Smart City software platform.
ABOUT IAN SIM
Ian Sim is the Co-Founder and CTO of Radical I/O. As a certified Enterprise Architect with 20 years of progressively-advanced technical leadership, Ian has a solid technical background, broad technology domain experience, and a proven ability to design both applications and multi-tiered, fault-tolerant, scalable systems.
Join our community by subscribing to Radical Ideas.