comscore
29 Feb, 2024 | Thursday
Trending : LaptopsAppsAutomobileHow ToOPPO India

Google Bard AI can now generate images from text

As Bard now uses Google's Gemini Pro large language model (LLM), it gains the capability to generate images from text using new Imagen 2 model.

Edited By: Shubham Verma

Published: Feb 02, 2024, 10:08 AM IST | Updated: Feb 02, 2024, 03:27 PM IST

Google Bard now uses Google Gemini Pro LLM to offer image generation.
Google Bard now uses Google Gemini Pro LLM to offer image generation.

Google has announced that its generative AI chatbot Bard AI can now generate images from text — a feature that its more popular rivals Bing AI and ChatGPT have been offering for months. The new image generation capability will be available in English in most countries for no extra fee. The only other platform currently offering free access to this feature is Bing AI (Microsoft Copilot), whereas ChatGPT Plus needs a paid subscription.

Users can ask Bard to generate images using Google’s new Imagen 2 model, which is “designed to balance quality and speed, delivering high-quality, photorealistic outputs.” The new capability emerges as Bard is now powered by Google’s Gemini Pro large language model (LLM), which natively supports image generation. However, according to The Verge, Bard was previously expected to use the Gemini Ultra LLM, but as it turns out this model is still in development.

Asking Bard to generate images based on your textual input is as simple as asking any question. Type what you want to generate into the text box on Bard and hit Enter. You will see multiple options as answers to your query. Photos generated by Bard AI look more realistic than what DALL-E can create, but they are still no match for ChatGPT Plus and Midjourney’s resulting photos. At the same time, both ChatGPT Plus and Midjourney rely on a paid subscription.

Bard uses SynthID to embed digitally identifiable watermarks into the pixels of generated images, so it becomes easy to tell if they were AI-generated. As AI fakes and deepfakes become more common — for instance the recent episode of Taylor Swift whose sexually explicit images flooded X, there is a call for regulation on what all AI should be allowed to do. At the centre of it is image generation. Google says there are technical and safety guardrails in Bard to stop people from generating AI images of well-known people and limit any scope of violent, offensive, or sexually explicit content.

Additionally, Bard is also getting support for more languages. Even though the image generation capability will work in English for now, Bard can respond to textual questions in more than 40 languages in over 230 countries. The double-check feature that lets you verify the answers with Google search results now works in over 40 languages.

Get latest Tech and Auto news from Techlusive on our WhatsApp Channel, Facebook, X (Twitter), Instagram and YouTube.

Author Name | Shubham Verma

STAY UPDATED WITH OUR NEWSLETTER

Select Language