Google Launches Gemini on Bard AI Chatbot — and It Beats ChatGPT in Blind Testing
Google announced today the integration of its next-generation artificial intelligence model, Gemini, into its Bard chatbot platform. Gemini is expected to enhance Bard’s performance and reasoning capabilities, allowing it to seamlessly understand various types of inputs such as images, text, code, audio, and video.
While some concerns were raised regarding potential delays due to language comprehension issues, Google reassured reporters during a press call that Gemini worked well across multiple languages. The new version of Bard, powered by Gemini, will initially be available only in English.
Bard was first introduced by Google as an experimental response to the success of ChatGPT eight months ago. Since its launch, Bard has undergone several updates, transitioning from LaMDA to PaLM 2, and now to Gemini.
Sissie Hsiao, the Vice President for Google Bard, revealed that Bard emerged as the most preferred free chatbot in blind evaluations conducted with third-party testers.
Gemini consists of three versions: Nano, designed for mobile devices; Pro, which powers Bard; and Ultra, slated for release next year to support a new Bard Advanced. Google claims that Gemini Pro performs on par with or exceeds the capabilities of GPT-3.5, the model behind the free version of ChatGPT. While Google declined to disclose a direct comparison between Gemini Ultra and GPT-4, it affirmed that Gemini Ultra outperforms all models out there on key benchmarks.
Demis Hassabis, CEO of Google DeepMind, highlighted that Gemini goes beyond the capabilities of most AI models by not only comprehending different types of text but also code, audio, images, and video.
Over time, Bard has been progressively enhanced with multimodal features, such as analyzing image contents, and gained extensions enabling it to check flights, review YouTube videos, and even check emails.
With the introduction of Gemini, Bard enters a new phase of development. Although the outward appearance may remain unchanged, the upgraded engine promises improved capabilities across various domains, comparable to the free version of ChatGPT.
Gemini Pro, built on a refined version of Gemini in English, enhances Bard’s reasoning, planning, and understanding abilities compared to the previous version based on PaLM 2. While its current release supports text-based prompts, upcoming updates will enable compatibility with other media types, as Gemini was purpose-built to be multimodal.
Google plans to launch a new version of its chatbot, Bard Advanced, next year. While details about the pricing model are yet to be revealed, it will be empowered by Gemini Ultra, the most powerful variant of the AI model. Leveraging the potent computational resources of Google’s network of data centers, Gemini Ultra aims to tackle highly complex tasks while rapidly understanding and acting on diverse information.
Google emphasized the importance of conducting additional safety checks and testing before unveiling Bard Advanced, aligning with its commitment to a responsible approach since Bard’s initial launch.
Sissie Hsiao expressed excitement about the advancements brought by Gemini, positioning Bard as the world’s foremost AI collaborator. She encouraged users to leverage these innovations in their creative pursuits, learning endeavors, and explorations.
Furthermore, Gemini is expected to integrate with Google Workspace through Duet AI early next year. Its integration will enhance the generation of detailed text and images within Google Docs, Sheets, and Slides. Google is also exploring the application of Gemini in Search, aiming to accelerate results and enable more complex queries.
As Google introduces Gemini on Bard, it harnesses the potential to revolutionize AI collaboration, infusing the chatbot with an unprecedented combination of capabilities. Users can now anticipate an enhanced and intelligent conversational experience while leveraging Gemini’s transformative powers across various media.
[Word Count: 600]
