A Secure Friend Recommendation Framework for Online Social Networks using OpenAI Embeddings

The popularity of Online Social Networks (OSNs), such as Facebook and Twitter, have led organizations to use OSN features to improve their business operations. Although OSNs amass billions of followers, privacy remains a pertinent concern for many users. A common functionality of OSNs is to facilitate friend recommendations (FR) without compromising user privacy. While users can add friends manually, most OSNs utilize different FR approaches that collect users profile data and employ linkage metrics to recommend new friends. In this paper, we propose an efficient and secure framework for creating friend recommendations. At the core of our framework, we use the OpenAI text embeddings to study the benefits of using new AI platforms and demonstrate its applicability to effectively address the FR problem in a privacy-preserving manner. Furthermore, our experimental results show that the proposed framework is superior in terms of accuracy, and efficiency, and also incurs minimal cost.