Mountain View: Google Bard, the artificial intelligence (AI)-powered chatbot, has undergone a significant update, introducing new features such as image generation and expanded language support.
The AI image generator within Google Bard can now create images from text inputs, leveraging the Imagen 2 model. The chatbot's image generation capabilities, powered by Imagen 2, promise high-quality, diverse, and realistic images from multiple paragraphs of text prompts.
However, it currently operates at a fixed resolution of 1536x1536.
To ensure the distinction between AI-generated and real photos, Google has implemented SynthID, a technology developed by its DeepMind division for watermarking and detecting AI-generated images. SynthID embeds a digital watermark into the pixels of an image, making it detectable for identification while remaining imperceptible to the human eye.
The update also brings global expansion, with Google Bard now available in over 230 countries and territories, compared to its previous availability in 170 countries. Furthermore, language support has been extended to over 40 languages, including Arabic, Bengali, Tamil, and Urdu, in contrast to its earlier support for English only.
Google has introduced a 'double-check' feature in all supported languages, enabling users to identify parts of the response sourced from the internet with citations, as well as parts without any reference. This feature aims to minimise instances of AI hallucinations, where the AI confidently provides incorrect responses.
The advancements in Google Bard come on the heels of Google's quarterly earnings call, where it was revealed that Google Bard Advanced, powered by Gemini Ultra, would offer a paid subscription.