New Delhi: Google’s Bard chatbot has enhanced its capabilities by introducing AI image generation. This feature puts Bard in competition with other platforms like ChatGPT Plus. The process is simple, allowing users to visit the site, input a prompt for image creation, and witness the magic happen. Notably, Bard’s service is free.
AI Image Generation With Room For Improvement
While Bard’s AI image generation is a step forward, it currently excels in creating overly simple photos. The tool is expected to improve over time.
Google’s Imagen 2 Text-To-Image Model
Now, Bard users can ask the chatbot to create pictures using Google’s Imagen 2 tool that turns text into images. Despite initial expectations of being powered by the more potent Gemini Ultra model, Bard is currently fueled by Google’s Gemini Pro large language model, with the Ultra model still in development.
Bard vs. ChatGPT Plus
Google positions Bard as a strong competitor to OpenAI’s ChatGPT Plus, which runs on GPT-4 and enables users to generate images through DALL-E 3 integration.
Previously, ChatGPT Plus had a slight advantage due to Bard’s lack of text-to-image features. However, with the integration of Imagen 2, Bard levels the playing field, and one notable difference is that Bard’s service is free, while ChatGPT Plus requires a subscription.
Google emphasizes responsibility in image creation with Bard’s capabilities. To make sure people know that the pictures created by the AI are not real, there’s a safety feature called watermarking.
It’s like a little mark that’s hidden in the pictures, showing that they were made by a computer and not by a person.
To make sure people use it responsibly, Google has put safety measures in Bard. These measures stop the tool from creating pictures of famous people or making images that are violent, mean, or show private stuff. This helps keep things positive and safe for everyone using the tool.
Google extends its generative AI exploration with ImageFX, an experimental photo tool powered by Imagen 2. Users can create images through simple text prompts.