3. November 2020
More camera data, less data privacy?
Nowadays, cameras are ubiquitous and leveraged for various topics, ranging from security to knowledge gain to the evolution of ai-driven technologies. The development of autonomous vehicles, high-definition maps, or smart retail analytics require large amounts of image and video data to be collected in public every day. The increasing number of cameras watching over public spaces infringes on the individual right for anonymity, a price the society paid for security in the past – and now also for innovation?
With increasing regulations worldwide like the EU’s GDPR, CCPA in the US, China’s CSL and APPI in Japan, companies, public entities and individuals are required to protect personal information, which includes biometric data in images and videos. While privacy regulations in different regions pose different legal bases for data collection and processing, they do have one in common: Consent. In the case of publicly recorded video data, however, it is not often not feasible to ask every data subject for consent. So, what can be done with “reasonable means” as GDPR mandates?
Various techniques to comply – anonymization & privacy by design
There are various technical and organizational means in order to comply (more) with relevant privacy regulations. Techniques to increase compliance range from meeting specific principles of data collection and processing (e.g. GDPR’s purpose limitation, data minimization and storage limitation) to specific technological means like encryption and decentralized processing to “classic” TOMs like privacy by default and by design. The most prominent examples of the latter one are anonymization and pseudonymization.
The GDPR defines anonymized data as ”information which does not relate to an identified or identifiable natural person or to personal data rendered anonymous in such a manner that the data subject is not or no longer identifiable“ (Recital 26). In other words: It should be impossible to derive insights on a discreet individual based on the anonymized data – even by the party that is responsible for the anonymization. When done properly, such data is by nature not personal and therefore not subject to data privacy regulations (Recital 26). Applied to image and video data, anonymization includes the redaction of PIIs like faces and bodies, as well as license plates. This way, personally identifiable information in images and videos is also protected against identification through facial or license plate recognition software.
AI-based redaction to anonymize image and video data
This can be done manually as well as fully automated. Especially for large data sets or numerous PIIs in single frames, manual work is tedious, slow and related to high costs. Automated solutions, on the other hand, are much faster and the easiest way for identity protection in camera data is de-identification software based on artificial intelligence. AI technologies are extremely efficient at detecting objects and can thus be used for automating image and video redaction in a fast and secure manner, without any human involved.
At brighter AI, we have steadily optimized these techniques for face and license plate redaction, and make them available to the broader public now via demo and free trial. brighter AI’s brighter Redact is the most advanced solution to protect personally identifiable information in image and video data. Here, we offer face blur and license plate redaction for image and video data – the anonymization solution is entirely automated and promises highest accuracies and quality standards. Furthermore, brighter AI’s proprietary Deep Natural Anonymization is now available as an additional feature in all paid plans – also in our SaaS-solution. This unique solution, in which the faces and license plates are replaced with synthetic image data with the help of generative AI, is compatible with the development of machine learning (ML) algorithms and analytics.
brighter AI’s brighter Redact is based on highest security standards and offers seamless integration (cloud to edge) as well as an interactive UI. It is the best product for privacy-compliant image and video processing – with privacy tech made in Germany. The benefits at a glance:
- Easy-to-use: Simple integration via cloud API or online user interface
- Precise: >99% anonymization accuracy
- Fast: Fully automatic, AI-powered anonymization software
- Secure: Hosted on MS Azure with TLS encrypted API
- AI-compatible: Deep Natural Anonymization (DNAT) enables analytics and machine learning compatibility
- Compliance: Anonymized datais by nature not personal and therefore not subject to data privacy regulations (GDPR, CCPA, CSL/PIS, etc.)