Microsoft adds voice, vision and o1 to Copilot
Microsoft is rolling out upgrades to its Copilot AI assistant, adding voice interaction, image analysis, and more sophisticated reasoning capabilities. These new features bring Copilot closer in line with the functionality of ChatGPT, on which it’s based.
The new “Copilot Voice” feature allows users to speak with the assistant, choosing from four different voice options.
For personalized content discovery, “Discover Cards” suggests topics and features based on a user’s interactions across Microsoft services.
“Copilot Daily” provides a morning briefing of news and weather, read aloud by the user’s preferred Copilot voice. Microsoft has partnered with news outlets including Reuters, Axel Springer, Hearst Magazine, and the Financial Times for this feature.
Ad
These new features will be rolled out across iOS, Android, web and Windows platforms. WhatsApp will also get Copilot integration. The voice feature will initially be available in English in select countries, while ‘Copilot Daily’ will launch in the US and UK.
Copilot reads websites and “thinks deeper”
In addition, Copilot is now integrated directly into the Microsoft Edge browser. Users can access it from the address bar to answer questions, summarize web pages, translate text, or rephrase sentences. Microsoft says it won’t use any personal data to train its AI systems.
Microsoft is also testing experimental features in Copilot Labs. These include “Copilot Vision,” which analyzes web pages in real time to answer questions about text and images, and “Think Deeper,” which provides more detailed answers to complex questions. The latter is likely to be powered by OpenAI’s new “o1” language model.
“Copilot Vision” and “Think Deeper” are initially only available to a limited number of Copilot Pro users in the USA and selected English-speaking countries.
The company is also enhancing its Paint application with AI-powered features. “Generative Fill” allows users to insert AI-generated elements into images based on text descriptions, while “Generative Erase” can remove objects from images, similar to Google Pixel’s Magic Eraser.
Recommendation
Windows search on Copilot Plus PCs is receiving an AI upgrade as well. A new feature called “Click to Do,” similar to Google’s “Circle to Search,” will allow users to search for images using words, even if they don’t appear in the photo or file name.
This feature will initially be available in File Explorer starting in November. Both new functions are exclusively available for Copilot Plus PCs.