By 2022, it’s predicted that 7 exabytes of video data will be transmitted by vehicles and public cameras per hour. This data is crucial for the development of autonomous driving infrastructure. But it also presents a major privacy protection concern.
This is where Brighter AI comes in. It’s a privacy tech start-up based in Berlin that was founded in 2017. It has developed a privacy solution for visual data called ‘Deep Natural Anonymization’ which helps companies to protect the identity of individuals and be compliant with increasing data privacy regulations worldwide, while collecting and processing camera data.
Auto Futures talked to Brighter AI’s Co-Founder and CEO, Marian Gläser, about why its solution is so necessary in the age of GDPR regulations, video streaming and human face recognition software.
“Deep Natural Anonymization is a SaaS or on-premise solution that generates artificial objects such as faces and license plates in order to replace original identifiers and therefore, protects peoples’ privacy. At the same time, in contrast to traditional anonymization methods like blurring, the data appears natural and can be used for analytics and machine learning,” says Gläser.
He adds: “This empowers companies to use publicly-recorded camera data for use cases such as the development of autonomous vehicles or driving monitoring while being compliant with privacy regulations.”
Gläser believes that, in the future, autonomous cars will stream video data out to help continuously improve their behaviour. Deep Natural Anonymization offers the ability to collect rich camera data, while protecting the privacy of pedestrians. Likewise, it enables driver monitoring without collecting a driver’s identity.
“The solution involves quite some magic and we cannot reveal all details. On a high level, I can say that our product pipeline involves up to eight neural networks, including generative adversarial networks. Regarding the faces, our models are trained on generating artificial objects that are significantly different and ensure cybersecurity compliance, but also maintain important attributes like age or eye gaze, which are critical in a variety of use cases,” says Gläser.
A Brighter Future
In 2018, Brighter AI was named as ‘Europe’s Hottest AI Start-up’ by none other than NVIDIA. In presenting the award to Brighter AI, NVIDIA’s CEO, Jensen Huang, said that the next wave of computing will be focused on such companies.
Gläser says: “We are already providing our solution to different European tier 1 suppliers and currently are in discussions with OEMs directly. One of our clients, a tier 1 supplier with approximately €20bn in revenues, aims to use data collection vehicles for autonomous driving in Asia. In this case, cybersecurity laws limit public data collection especially for foreign companies. For this client, we are deploying our natural anonymization directly on the test vehicle and solving their legal headaches.”
At present, Brighter AI has a team of twelve employees based in Germany. It hopes to more than double that number by the end of the year. It also has plans to expand internationally.
Gläser tells us: “We are currently expanding into more verticals such as retail and public transportation. Here, we already started a project with Europe’s largest public transportation provider. Also, we see privacy regulations evolving worldwide, with GDPR being the global role model. In 2020, the U.S will get new state laws such as the California Consumer Privacy Act. Therefore, we will push our internationalization and plan to have a U.S office next year.”
A Clearer Vision
As for the future of big data and autonomous driving, Gläser predicts: “In the future, all cameras will be connected to the cloud where data will be processed for business analytics. This trend can already be seen in the consumer space where solutions like the Amazon Cloud Cam stream video footage directly into the cloud. Edge computing will be relevant for pre-processing, including the anonymization of personal identifiers. But the main analytics will take place in the cloud.”
He concludes: “Data sharing for autonomous vehicles, cities, companies, insurances and more will be crucial. And this requires privacy by design, especially for cameras. No one wants to live in a ‘George Orwell 1984 society’. Therefore, privacy will be a core part of every product in the future.”