Privacy in Responsible AI: Approaches to Facial Recognition from Cloud Providers
Anna Elivanova
TL;DR
The study tackles privacy in cloud-based facial recognition within Responsible AI by performing a comparative analysis of Microsoft Azure, AWS, and Google Cloud. It synthesizes official documentation, guidelines, and policy discussions to map each provider's privacy measures, governance, and the shared responsibility model. Key contributions include detailing Microsoft’s restricted, case-based access and explicit responsible-use guidelines, AWS’s broad API availability with heightened customer responsibility, and Google’s detection-focused approach with deprecation of celebrity recognition, along with unified data-protection tools and compliance resources. The findings offer practical insights for developers and businesses to implement privacy-preserving cloud FRT and underscore the need for more standardized privacy practices across cloud providers to reduce ambiguity and risk.
Abstract
As the use of facial recognition technology is expanding in different domains, ensuring its responsible use is gaining more importance. This paper conducts a comprehensive literature review of existing studies on facial recognition technology from the perspective of privacy, which is one of the key Responsible AI principles. Cloud providers, such as Microsoft, AWS, and Google, are at the forefront of delivering facial-related technology services, but their approaches to responsible use of these technologies vary significantly. This paper compares how these cloud giants implement the privacy principle into their facial recognition and detection services. By analysing their approaches, it identifies both common practices and notable differences. The results of this research will be valuable for developers and businesses by providing them insights into best practices of three major companies for integration responsible AI, particularly privacy, into their cloud-based facial recognition technologies.