{"data":{"id":2199,"lang":"en","title":"Google Unveils HEIR, an Open-Source Compiler for AI on Encrypted Data","summary":"Google's new open-source HEIR compiler lets AI models run inference on encrypted data without decrypting it, aiming to make private AI practical for non-experts.","category":"ai","category_label":"AI","published_at":"2026-08-16T14:01:16Z","updated_at":"2026-08-16T14:01:16Z","url":"https://techtools.cz/tools/news/article/2199-google-unveils-heir-an-open-source-compiler-for-ai-on-encrypted-data","api_url":"https://techtools.cz/news-api/v1/articles/2199?lang=en","markdown_url":"https://techtools.cz/news-api/v1/articles/2199.md?lang=en","image_url":"https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Untitled_design_10_p3wN1dg.width-1200.format-webp.webp","word_count":145,"reading_time_minutes":1,"sources":[{"url":"https://blog.google/security/how-google-is-making-private-ai-practical-with-homomorphic-encryption/","domain":"blog.google"}],"translation":{"lang":"cs","id":2272,"title":"Google představil HEIR, open-source kompilátor pro AI nad šifrovanými daty","url":"https://techtools.cz/tools/news/clanek/2272-google-predstavil-heir-open-source-kompilator-pro-ai-nad-sifrovanymi-daty","api_url":"https://techtools.cz/news-api/v1/articles/2272?lang=cs"},"content":"Google announced HEIR (Homomorphic Encryption Intermediate Representation) on August 14, an open-source compiler toolchain that lets AI models run inference directly on encrypted data, without ever decrypting it on the server.\n\nThe tool targets a long-standing trade-off: strong encryption protects user data but usually blocks providers from offering data-dependent features like recommendations or fraud checks. HEIR converts pre-trained models so they operate on encrypted inputs instead, with Google describing the goal as a \"one-click\" path for non-experts to add encrypted inference to production apps. Google demonstrated the approach on four applications: a recommendation model, credit-card fraud detection, network intrusion detection, and hotword detection for voice assistants.\n\nThe project already has hardware-accelerator partners including Belfort Labs, Niobium, Cornami, and Optalysys, plus academic collaborators such as Georgia Tech, Carnegie Mellon, and Tsinghua University. Four peer-reviewed papers have been built on HEIR so far, with more in preparation."}}