Google’s Bard chatbot is getting a new feature – AI image generation. This catches up with a similar feature that the rival ChatGPT Plus had for quite a while.
Users can now ask Bard to create pictures using Google’s Imagen 2 text-to-image model. Bard, now using Google’s Gemini Pro large language model, was expected to use the more powerful Gemini Ultra model, but that one is still being worked on.
Google Bard Text to Image Remains Free to Use 
Google has been positioning Bard as a competitor to OpenAI’s ChatGPT Plus, which uses GPT-4 and allows users to make images with DALL-E 3 integration. Both chatbots work well, but Bard didn’t have text-to-image features before, giving ChatGPT Plus a bit of an advantage. The updated Bard with Imagen 2 is available to all users for free, unlike ChatGPT Plus, which needs a paid subscription.
AI image generation faced criticism recently after fake explicit photos of Taylor Swift appeared online. While it’s unclear where these images actually came from, tech companies are now tightening security on their image generating platforms.
Gizchina News of the week
Google emphasized responsibility in the design of Bard’s image capability. The company is adding a watermarking feature to the AI-generated images, embedded in the pixels, to indicate their artificial origin. Google has taken steps to implement technical measures and safety precautions to prevent the generation of images involving well-known individuals and to restrict violent, offensive, or sexually explicit content.
The image generation feature is not exclusive to Bard; Google introduced a new experimental photo tool called ImageFX, powered by Imagen 2. ImageFX enables users to create images using simple text prompts, allowing people to explore generative AI, with feedback sent directly to Google engineers. Similar to Bard, all photos created through ImageFX will be marked by SynthID and adhere to Google’s AI principles and technical guidelines. ImageFX joins Google’s other AI creation tools, MusicFX and TextFX, which the company launched last year and are available for testing on Google Labs.
Google Bard Expands to More Countries 
Bard is receiving several other updates. While image generation is currently available only in English, the chatbot has expanded to support over 40 languages across more than 230 countries. The previous version was only available in English and accessible in 170 countries. Supported languages now include Arabic, Bengali, Tamil, Urdu, and others.
It is worth noting that, Google also extended the double-check feature to 40 languages, allowing users to assess responses by automatically searching the internet to verify or contradict the provided content.
