Top news
Investing in the public benefit OfficialAnthropic NewsOct 18Popular Musk Pledges SpaceX Shares to ‘Every Kid in America’ WebBloomberg TechnologyOct 19Popular Websites are closing their doors to AI agents WebThe Verge AIOct 17 This is the new ChatGPT Voice, powered by GPT Live OfficialOpenAI on YouTubeOct 18 How AI giants will reshape global philanthropy WebAxios AIOct 19 Gemini 4 Argon Really Surprised Me WebTwo Minute PapersOct 19 Google Gemini 4 agents just beat Minecraft with no human intervention WebEngadget AIOct 19 Who could it be? 🧐 OfficialGoogle on YouTubeOct 18 Anthropic gives outside AI watchdog “employee-like” access WebArs Technica AIOct 18 What are self-replicating ‘AI worms’, and why is Anthropic worried about them? WebThe Guardian AIOct 18
GlossaryModelsIntermediate

Code-switching

Also written: Context switching, Multilingual code-switching

Switching between two or more languages or language varieties within a single conversation.

In plain terms

Code-switching is when someone mixes languages or dialects while talking. It's common for people who speak more than one language.

For example, a person might start a sentence in English and finish it in Spanish, or use a word from one language that doesn't have a direct translation in another. AI models are getting better at understanding and even using code-switching themselves, especially in tasks like translation or transcription.

The expert version

Code-switching refers to the practice of alternating between two or more languages, dialects, or language varieties within a single discourse. This phenomenon is a natural aspect of bilingual and multilingual communication. In AI, models capable of handling code-switching can process and generate text that reflects this linguistic behavior, which is crucial for applications dealing with diverse user inputs or multilingual content.

Why it matters now

Recent advancements in speech-to-text models, like Google's Gemini 3.5 Transcribe, highlight the growing importance of multilingual capabilities. These models are being developed to handle diverse language inputs and switch between languages seamlessly, reflecting a trend towards more versatile AI communication tools. This focus on accurate and low-latency transcription across multiple languages is crucial for global applications.

Where it shows up 10 items, newest first
How each company uses it
Google Gemini
Google Gemini's recent updates to its Gemini 3.5 Transcribe model emphasize its ability to support multiple languages and accurately convert speech to text, showcasing its progress in multilingual AI.
OpenAI ChatGPT
OpenAI's GPT Transcribe model is noted for its advanced speech-to-text features, including multilingual support, indicating its focus on handling diverse language inputs.
Related terms

MultilingualismLanguage identificationSpeech recognitionNatural language understandingDialectTranslation

Often confused with
Code-mixing
Often used interchangeably with code-switching, though some linguis
Language transfer
Refers to the influence of a speaker's native language on the
Pidgin
A simplified language that develops as a means of communication between
Origin

The term 'code-switching' originated in linguistics to describe the phenomenon observed in bilingual and multilingual communities. It has been studied extensively in sociolinguistics since the mid-20th century. Source ↗

Try this about 5 minutes

Try speaking to an AI chatbot, like ChatGPT or Google Gemini, and intentionally switch between two languages within the same sentence or question. Observe if the AI understands your request and responds appropriately, potentially even acknowledging the language switch.

Try it: Playbook entries that use thisPlaybook →

The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.

Questions people ask

Is code-switching a sign of poor language skills?
No, code-switching is a natural and often sophisticated linguistic practice used by fluent bilingual and multilingual speakers.
Can AI models truly understand code-switching?
Modern AI models are increasingly capable of recognizing and processing code-switched input, though nuances can still be challenging.

First seen on Subvolts Nov 25, 2025. Updated Oct 15, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.