ASR
Also written: Speech-to-text, Speech recognition system, Voice to text, Automatic Speech Recognition
Automatic Speech Recognition (ASR) converts spoken words into written text.
Automatic Speech Recognition, or ASR, is technology that allows computers to understand and write down what people say. Think of it like a super-powered version of voice dictation on your phone or computer.
This technology is what makes features like voice assistants (like asking Siri or Google Assistant a question) and automatic captioning on videos possible. It takes the sound waves of speech and turns them into readable text.
Automatic Speech Recognition (ASR) is the process by which spoken language is converted into a machine-readable text format. It typically involves acoustic modeling, which maps audio signals to phonetic units, and language modeling, which predicts the likelihood of word sequences.
Modern ASR systems often leverage deep neural networks, particularly recurrent neural networks (RNNs) and transformer architectures, to achieve high accuracy. Tradeoffs exist between real-time processing, computational cost, and accuracy across different acoustic environments and speaker variations.
Frontier AI labs are improving Automatic Speech Recognition (ASR) to make AI interactions more natural and accessible. Recent advancements focus on supporting more languages and enabling real-time, full-duplex voice conversations, moving beyond simple transcription to more intuitive communication.
- These execs think voice AI hasn’t reached its ChatGPT moment yetOpenAI · TechCrunch AI ·
- ChatGPT Intelligent UI 🧠, Claude Haiku 5.5 ⚡, Grok Bot 3rd party models 🤖Anthropic · TLDR AI ·
- Build more natural voice experiences with GPT‑Live‑1 in the APIOpenAI · OpenAI News ·
- human-level robustness and accuracy on English speech recognitionOpenAI · OpenAI Research ·
- Introducing Meta Omnilingual Automatic Speech Recognition | Transcription for 1,600+ languagesMeta · AI at Meta on YouTube ·
- OpenAI Data PartnershipsOpenAI · OpenAI News ·
- ChatGPT can now see, hear, and speakOpenAI · OpenAI News ·
- Anthropic Claude
- Anthropic's Claude models are part of the ongoing development in AI that can process and understand spoken language.
- Meta Meta AI
- Meta AI developed an Omnilingual ASR system that supports over 1,600 languages, significantly expanding speech recognition capabilities.
- OpenAI ChatGPT
- OpenAI has integrated ASR into ChatGPT, allowing it to understand voice input, and released the Whisper model for robust speech recognition and translation.
- [object Object]
- [object Object]
The concept of converting speech to text has roots in early signal processing and linguistics research dating back to the mid-20th century. The term 'Automatic Speech Recognition' became prevalent as computational power increased. Source ↗
Use the voice typing feature in Google Docs or your smartphone's keyboard. Speak a few sentences and see how accurately the text appears on the screen. Then, try reading the same sentences in a noisy environment or with a different accent to observe potential challenges.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- How accurate is ASR?
- Accuracy varies greatly depending on the model, audio quality, background noise, and accents. Modern systems can achieve very high accuracy, often over 90% for clear speech in quiet environments.
- What are the main challenges for ASR?
- Key challenges include handling background noise, different accents and dialects, overlapping speech, and variations in speaking speed and style.
First seen on Subvolts Sep 21, 2022. Updated Oct 18, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.