Google offers Bard access to US, UK users with 'more coming soon'

Google has announced the launch of Bard, an experimental AI tool that allows users to collaborate with generative AI to boost productivity and creativity. 

Bard is based on a research large language model (LLM), which selects words likely to come next when given a prompt, and generates a response by selecting words one at a time. Bard's LLM is a lightweight and optimised version of LaMDA, grounded in Google's understanding of quality information.

Bard is designed to help users achieve their goals, from tips on reading more books to outlining a blog post. Although LLMs are not without their faults, including reflecting real-world biases and presenting inaccurate information, Google claims it has built-in guardrails, such as capping the number of exchanges in a dialogue, to keep interactions helpful and on topic. Users are also given the choice of a few different drafts of Bard's response to pick the best starting point.

Google sees Bard as a complementary experience to Google Search, allowing users to easily visit Search to check its responses or explore sources across the web. The company plans to thoughtfully integrate LLMs into Search in a deeper way in the future.

In its blog post, Google said it is guided by its AI Principles in building Bard responsibly, and is using human feedback and evaluation to improve the system. The company plans to add capabilities to Bard, including coding, more languages, and multimodal experiences, and will continue to learn alongside users.

Bard has begun rolling out access in the US and UK, with plans to expand to more countries and languages over time. Users can sign up to try Bard at bard.google.com.