The provided text discusses a recent development in artificial intelligence (AI) by Apple, which uses homomorphic encryption to process data in a privacy-preserving manner. 

Homomorphic encryption allows calculations to be performed on encrypted data without decrypting it, meaning that servers and AI models can process information without directly accessing the original data.

This approach, which Apple calls its "HTTPS moment" for AI, addresses privacy concerns associated with data processing and storage. By combining private information retrieval (PIR) and private nearest neighbor search (PNNS), users can submit queries and have their data analyzed while ensuring their privacy. This has significant implications for AI applications that involve sensitive data, such as search, image recognition, and other functions.

The article's author, Enrique Dans, highlights the importance of this development for alleviating concerns regarding data security and user privacy in the context of AI.

This article is also available in English on my Medium page, «Is AI about to have its HTTPS moment?»