© IE Online Media Services Pvt Ltd
Tags:
Journalism of Courage
Google’s conversational AI chatbot Bard just got an upgrade with some nifty new capabilities – most notably the ability to generate images from text prompts and double-check its own responses in more languages. Let’s take a deeper look.
The most eye-catching new feature is undoubtedly image generation. Users can now describe an image to Bard in natural language, and it will conjure up a custom visual. This leverages Google’s updated Imagen 2 model to produce high-quality, photorealistic images on demand.
So if you tell Bard “create an image of a dog riding a surfboard,” it can instantly generate the whimsical scene. The possibilities are endless – you can use it to mock up designs, illustrate concepts, or just for fun.
Google says Imagen 2 balances quality and speed, enabling quick generation times. And the images have built-in digital watermarks via SynthID technology. This helps differentiate AI creations from human-made artwork. The company also the model has technical safeguards to limit harmful content. The training data is vetted to avoid violent, offensive or explicit material. And filters block the creation of specific people’s likenesses.
Bard is also expanding its handy double-check feature to over 40 languages. This lets users click an icon to have Bard corroborate its responses against reputable online sources. Previously only in English, double-check now works in everything from Hindi to Portuguese. This helps Bard provide factual information you can trust in more regions.
In addition to the visuals, Bard is getting smarter internally. Google is expanding the more advanced Gemini Pro capabilities to over 40 languages in over 230 countries. Gemini Pro boosts Bard’s understanding, reasoning, summarizing, and coding abilities. This allows for faster, more nuanced conversations. Google claims early analysis shows Bard with Gemini Pro is one of the top-performing AIs available.
With these upgrades, Bard continues advancing as a versatile AI assistant. The image generation unlocks new creative possibilities, while broader language support for double-check improves reliability.
