Google has announced an exciting update for its generative AI chatbot, Bard. Powered by Gemini, Google’s latest and most advanced AI model, Bard will now possess enhanced reasoning, planning, understanding, and other capabilities. With Gemini available in three sizes – Ultra, Pro, and Nano – Bard can be deployed across a range of devices, from mobile to data centers.
Key Takeaway
Google’s AI chatbot, Bard, will receive a significant upgrade with the integration of Gemini, the company’s most advanced AI model. This update will empower Bard with improved reasoning, planning, understanding, and other capabilities.
Rollout Approach and Availability
The deployment of Gemini to Bard will occur in two phases. Initially, Bard will be upgraded with a tailored version of Gemini Pro. In the next year, Google will introduce Bard Advanced, which will grant users access to the best AI model, starting with Gemini Ultra. The version of Bard equipped with Gemini Pro will be available in English across more than 170 countries and territories globally. Additionally, Google plans to expand its availability to other languages and countries, including the EU and U.K.
Benchmark Performance and Comparison
Prior to its public release, Gemini Pro underwent a series of industry-standard benchmarks. Google claims that Gemini outperformed GPT-3.5 in six out of eight benchmarks, including the key standards for measuring large AI models, such as massive multitask language understanding tasks (MMLU) and grade school math reasoning (GSM8K). However, it’s worth noting that GPT-3.5 is over a year old, bringing a sense of catch-up to this launch rather than an outright outperformance.
Improved Capabilities and Applications
The upgrade aims to enhance Bard’s capabilities in understanding and summarizing content, reasoning, brainstorming, writing, and planning. Sissie Hsiao, Vice President and General Manager of Assistant and Bard at Google, referred to this update as the most substantial quality improvement since Bard’s launch.
Gemini Pro will initially power text-based prompts in Bard, but the plan is to expand support to multimodal inputs, including texts, images, and other modalities. Bard Advanced, set for release in 2024, will usher in a new experience using Gemini Ultra, an AI model proficient in understanding and acting upon different types of information, such as text, images, audio, video, and code. Google claims that Gemini Ultra can generate high-quality code in popular programming languages and possesses multimodal reasoning capabilities. This upcoming upgrade promises to be the most anticipated one yet.
Future Plans and Safety Measures
Google plans to launch a trusted tester program for Bard Advanced in early 2023 before expanding availability to a broader userbase. Additionally, the company will conduct additional safety checks on Bard Advanced before its official launch.
Continued Enhancements to Bard
Since its debut just eight months ago, Bard has undergone various improvements. Recent updates have introduced features enabling Bard to answer questions about YouTube videos, access users’ Google apps like Gmail, Docs, and Drive, as well as utilize other Google services such as Google Flights and hotels. Notably, Bard can double-check its answers to identify instances of providing responses based on false information or “hallucinating.”
“Now with Gemini, we’re one step closer to bringing you the best AI collaborator in the world,” stated Sissie Hsiao, highlighting Google’s commitment to continuously enhancing Bard’s capabilities.
Bard with Gemini is available for use starting today.
