# How Smart Speakers Recognize Who Is Talking > This patent describes how a speech-controlled device identifies a specific user from their voice to deliver personalized content or deny requests based on recognition confidence. - **Patent:** US 12626703 - **Original title:** User recognition for speech processing systems - **Owner:** Amazon Technologies - **Granted:** 2026 - **Status:** Active - **Times cited:** 0 - **Field:** consumer_electronics, software, telecommunications, ai_ml ## What it does This patent describes a system where a device, like a smart speaker, captures spoken audio (Claim 1). A server then analyzes this audio to figure out who spoke it, matching the voice to a "user identifier" (Claim 1). Once the speaker is identified, the system processes their request and provides "first content" that is "customized based at least in part on the user identifier" (Claim 1). For example, if a child asks for music, the system might recognize their voice and play music from their specific playlist. The patent also covers scenarios where a "user recognition condition" must be met (Claim 2). If the system can't confidently identify the user for a sensitive request, it might "decline to execute the second request" (Claim 3) or provide generic content instead of customized content (Claim 4). It can even use face recognition alongside voice for stronger identification (Claim 10). ## What it does NOT cover - Does not cover speech processing systems that do not attempt to identify the specific user speaking. - Does not cover systems that customize content without using the user's voice for identification. - Does not cover systems that identify users solely through non-biometric means, like a login or explicit command. - Does not cover systems that only use image data for user recognition without also processing audio data (Claim 10 requires both for enhanced recognition). - Does not cover systems that always provide customized content regardless of user recognition confidence. ## The clever bit The clever part is linking user identification directly to content customization and access control, especially by considering "user recognition conditions" (Claim 2) and potentially using "most stringent" confidence thresholds (Abstract). This allows the system to intelligently decide when to personalize content, when to provide generic content, or when to deny a request entirely based on how sure it is about who is speaking. ## Real-world examples 1. Amazon Echo devices (Alexa Voice Profiles) 2. Google Nest speakers (Voice Match) 3. Apple HomePod (Recognize My Voice) 4. Smart TVs with voice control 5. In-car voice assistants ## Why it matters This technology is fundamental to modern voice assistants and smart home devices. It allows a single device to serve multiple family members with personalized experiences, like playing their specific music, accessing their calendars, or making purchases from their accounts. Without robust user recognition, voice assistants would be less useful and potentially insecure for sensitive tasks. Amazon, the assignee, heavily relies on such systems for its Alexa-enabled devices. ## Frequently asked questions ### What does How Smart Speakers Recognize Who Is Talking cover? This patent describes how a speech-controlled device identifies a specific user from their voice to deliver personalized content or deny requests based on recognition confidence. ### Who owns patent US 12626703? Amazon Technologies owns this patent, granted in 2026. ### When does this patent expire? This patent is expected to expire on May 14, 2044, when the invention enters the public domain. ### What problem does this patent solve? This technology is fundamental to modern voice assistants and smart home devices. It allows a single device to serve multiple family members with personalized experiences, like playing their specific music, accessing their calendars, or making purchases from their accounts. Without robust user recognition, voice assistants would be less useful and potentially insecure for sensitive tasks. Amazon, the assignee, heavily relies on such systems for its Alexa-enabled devices. ### What does this patent NOT cover? Does not cover speech processing systems that do not attempt to identify the specific user speaking. **Full plain-English explainer:** https://patentbrief.org/patent/us/12626703/user-recognition-for-speech-processing-systems **Original patent:** https://patents.google.com/patent/US12626703 --- _Source: PatentBrief — https://patentbrief.org. Patent facts are from public records; the plain-English explanation is PatentBrief's._ ## Related patents Semantically similar inventions in the PatentBrief corpus: - [How Voice Assistants Recognize Who is Talking](https://patentbrief.org/patent/us/11514901/anchored-speech-detection-and-speech-recognition) — This patent describes a system that uses a recorded voice sample to identify if new speech comes from the same person, allowing voice assistants to only respond to specific users. - [How Voice Assistants Distinguish Speakers After a Wakeword](https://patentbrief.org/patent/us/10923111/speech-detection-and-speech-recognition) — This patent describes a method for voice assistants to better understand who is speaking after a 'wakeword' by comparing voice characteristics to improve speech recognition accuracy, especially in multi-speaker environments. - [How Sonos Speakers Use Personalized Wake Words to Recognize Different Users](https://patentbrief.org/patent/us/9965247/icloud-drive) — A system that lets multiple people control a shared speaker by using unique voice-trigger words to link their specific music accounts and preferences. - [How Wearable Devices Use Location and Biometrics to Know Who You Are](https://patentbrief.org/patent/us/20220189458/speech-based-user-recognition) — This patent describes how a computer system can figure out who is using a wearable device by combining its location data with biometric information like a face scan or voice recording. - [How Voice Assistants Change Their Speech Based on How You Talk](https://patentbrief.org/patent/us/10276149/dynamic-text-to-speech-output) — This patent describes a system where a voice-controlled device adjusts its text-to-speech output characteristics, like speed or tone, based on the user's speaking habits or current situation, making responses feel more natural and personalized.