On February 1, Google Bard underwent a major upgrade, introducing several new features to enhance its capabilities as an AI-powered chatbot. The most noteworthy enhancement is the incorporation of an AI-based image generation feature. It allows you to enable the chatbot to now create images based on textual inputs. It’s worth noting that, as of now, Google Bard can’t produce image-to-image outputs.
In addition to the image generation feature, Google expanded the availability of Bard to more than 230 countries and territories, emphasizing its commitment to a global user base. Furthermore, the language support has been extended to over 40 languages, a significant expansion from its previous availability in 170 countries and English-only language support. This move positions Google Bard as a more accessible and inclusive tool for users around the world.
The announcement of this update closely followed Google’s quarterly earnings call, where it was revealed that Google Bard Advanced, powered by Gemini Ultra, will soon introduce a paid subscription model. This strategic move aligns with the growing trend in the industry of offering advanced AI features through subscription services.
The introduction of the AI image generator in Google Bard is a notable development, catching up with competitors such as OpenAI’s ChatGPT Plus, Microsoft Copilot, and Baidu’s Ernie Bot, which had already incorporated this feature. The image generation capabilities in Google Bard are attributed to the Imagen 2 model, the same model powering other Google AI products like ImageFX and Vertex AI, currently in the testing phases.
The AI image generator in Google Bard allows for the creation of images from text inputs, even from prompts spanning multiple paragraphs. Google asserts that the generated images are of high quality, diverse, and realistic. In our own testing, we found that the AI model performed reasonably well, producing images that aligned with the provided prompts.
However, it’s important to note that the generated images are fixed at a resolution of 1536×1536 and are distinguishable as digitally created in most cases. Notably, Google Bard refuses requests to generate images of real-life people, a measure likely taken to mitigate the risks associated with deepfake technology.
In addition to the image generation capabilities, Google has implemented SynthID to label the images created by Google Bard as AI-generated. SynthID, developed by Google’s DeepMind division, is a tool for watermarking and detecting AI-generated images, introduced in August 2023. This technology embeds a digital watermark into the pixels of an image, invisible to the human eye but detectable for identification.
Beyond these enhancements, Google Bard now boasts an extended language support of over 40 languages, including Arabic, Bengali, Tamil, and Urdu. The ‘double-check’ feature has also been integrated across all supported languages, offering users the ability to identify the parts of a response derived from referenced sources on the internet, along with citations.
This feature serves to minimize instances of AI-generated responses with incorrect information, addressing the phenomenon known as AI hallucinations. Accessible by clicking the G icon beneath the generated response, the ‘double-check’ feature enhances transparency and reliability in user interactions with Google Bard.
Read more about:
If you have any other questions related to Google’s new AI Image Generation tool, let me know in the comment section. Thank you for reading. Have a nice day.
Did you enjoy this post? Share it!