Google has expanded its AI chatbot, Bard, powered by Gemini Pro, to support over 40 languages in more than 230 countries. The free chatbot allows users to create images in English in most countries.
With Gemini Pro enhancing Bard’s understanding, reasoning, and coding abilities, the competition between top AI chatbots is tightening. Although Gemini Pro doesn’t outclass GPT-3.5 Turbo in benchmark tests, the results showcase a competitive edge.
One advantage Bard offers over OpenAI’s free ChatGPT platform is its real-time internet access, as ChatGPT’s knowledge only extends until September 2019. Bard also provides a feature called ‘Double Check’, designed to verify the truthfulness of its generated responses by cross-referencing with internet content.
There is speculation that Google is planning to rename Bard to reflect the Gemini brand, as suggested by a leaked change log.
One of Bard’s key functionalities is text-to-image generation for English-speaking users in most countries, powered by Google’s Imagen 2 model. The images may not be comparable to Midjourney v6 or DALL-E, but considering the feature is free and offers a step up from the text-only ChatGPT, they are remarkably impressive.
Google has put measures in place to restrict violent, offensive, or sexually explicit content, including the generation of named persons’ images. Bard’s alignment to these content restrictions may sometimes prevent it from generating certain images. Images created by Bard will carry a digital watermark added through SynthID, following Google’s AI Principles.
Bard, with its multimodal functionality and integration with other Google products like Gmail and Calendar, offers more than just a chatbot experience. It might entice more ChatGPT users with these enriched features. It remains to be seen how OpenAI will respond to these new features and the performance boost given to Bard by Google.
