Zoom Introduces AI Companion 2.0 with Enhanced Features
Karnataka News
At its annual Zoomtopia event on Wednesday, Zoom unveiled exciting new artificial intelligence (AI) features for its platform. The highlight is the AI Companion 2.0, an upgraded version of its AI assistant that can now perform a wider range of tasks. This assistant is part of Zoom Workplace, the company’s AI-driven collaboration platform, and can now access features like Zoom Mail and Zoom Tasks.
Additionally, Zoom introduced a custom AI avatar feature. This allows users to create an AI clone of themselves by recording a video. The AI avatar can then speak in a voice similar to the user’s, making virtual interactions more personalized.
Read More: Google Photos Might Introduce AI Image Credits to Help Users Identify Deepfakes
Zoom AI Companion 2.0 Introduced
In a recent newsroom post, the company announced a new AI assistant for its Zoom Workplace platform. This assistant is similar to the Gemini assistant found in Android smartphones and the Copilot feature in AI-enabled PCs. The AI Companion serves as the central hub for many of the new AI features introduced by the company, allowing users to summarize documents, generate text, and much more.
With the second version of the AI assistant, users can now access prompt suggestions in the side panel across all Zoom platforms. Additionally, users can enhance their context by linking their Gmail, Microsoft Outlook, and other similar apps. The assistant can summarize unread messages in Zoom Team Chat channels and recap email threads in Zoom Mail.
During Zoom calls, the AI assistant can answer questions beyond the meeting context by searching the web. Participants can ask general knowledge or current affairs questions while in a meeting. Furthermore, AI Companion 2.0 can respond to inquiries about meeting content even after the meeting ends. Users can also refer back to their conversations with the AI after the meeting is over.
Zoom Introduces Custom AI Avatar
Another exciting feature introduced at the event is Custom Avatars for Zoom Clips. This allows users to record a video clip that an AI model processes to create an avatar. The avatar will only display the upper body, including the head, shoulders, and upper arms. It can also produce a voice that sounds similar to the user. If a user provides a text script, the AI avatar can speak and lip-sync to it.
Although this feature was showcased at the Zoomtopia event, it won’t be available to users until 2025.
Google Gemini Updated With Imagen 3 AI Model for Image Generation, Available to All Users
On Wednesday, Google announced a major upgrade for Gemini, its proprietary artificial intelligence (AI) model. The upgrade includes a new image generation feature that will now be powered by the Imagen 3 AI model, which is Google’s most advanced image generation technology to date. This enhancement will be available to all users of the Gemini app and is also being extended to the API version of Gemini, allowing developers to create applications and experiences that utilize this new capability.
Check out more latest App News
Gemini Users Get Access to Imagen 3 AI Model
In a post on X (formerly Twitter), the official Google Gemini App account announced that all users, including those on the free tier, will now have the ability to generate images using the Imagen 3 AI model. The post emphasized that Imagen 3 offers high photorealism, improved adherence to prompts, and minimizes unwanted elements in images.
Staff members at Karnataka News Express confirmed that the Gemini app is indeed using Imagen 3 for image generation. To evaluate its capabilities and compare it with Meta AI, they tested both chatbots with the same prompt: “Draw an image of a golden retriever dog sitting on a train berth, looking out through the window at the Alps. The train has a wooden interior, and the seats are green. All other passengers on the train are animals, and one human conductor is checking for tickets.”
The images generated by both AI models can be seen above. While neither model fully incorporated all the elements from the prompt, Gemini managed to include more details than Meta AI. Additionally, images produced by Meta AI have a resolution of 1280 x 1280, whereas Imagen 3 generates images at a higher resolution of 2048 x 2048.
Imagen 3 offers a variety of styles, including photorealistic images, textured oil paintings, and claymation scenes. Users can also specify that images be styled as if taken with certain cameras, such as a Nikon DSLR, GoPro, or wide-angle lens.
To mitigate the risk of deepfakes, Google has integrated safeguards into the AI model. Every generated image is watermarked with SynthID, a technology that embeds an invisible AI label within the image’s pixels. This watermark cannot be cropped out or removed, and it remains visible even in screenshots.
