Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More
The landscape of AI chatbots just became a bit more competitive. Google announced today that it’s opening access to Bard, the company’s experimental text-based service that lets you collaborate with generative AI. The company will slowly roll out access to the chatbot starting with the U.S. and U.K. markets and will expand to more countries and languages over time.
Bard, a conversational AI chatbot comparable to ChatGPT, Claude and Bing Chat, is built on Google’s Language Model for Dialogue Applications (LaMDA), which was first introduced in 2021. The new chatbot aims to emulate human-like conversations by utilizing natural language processing and machine learning to generate realistic and helpful responses to user queries.
>>Follow VentureBeat’s ongoing generative AI coverage<<
In its surprise release, Google acknowledges that large language models (LLMs) may have their shortcomings and that users should approach Bard with caution. “For instance, because they learn from a wide range of information that reflects real-world biases and stereotypes, those sometimes show up in their outputs. And they can provide inaccurate, misleading or false information while presenting it confidently.”
The company also includes a warning at the bottom of the chatbot text input box, saying Bard may display inaccurate or offensive information that doesn’t represent the company’s views.
Bard functions similarly to ChatGPT, Claude, Bing Chat and other AI chatbots. Users can input text, and Bard will generate a fitting, often surprisingly helpful reply. A distinctive feature of Bard is that it presents users with several drafts of its response, allowing them to choose the most suitable starting point for their request. If a completely new response is desired, users can prompt Bard again. This is reminiscent of the creative, balanced and precise options provided by Bing Chat.
A noteworthy point in the announcement is that Bard serves as a direct interface to LLMs and is intended to complement Google Search, according to the company. Bard also incorporates a “Google it” button, which redirects users to a relevant Google Search on some queries.
Throughout the announcement post, Google repeatedly emphasizes that “Bard is an experiment.” The company dedicates an entire section to explaining how Bard is guided by its AI Principles and how it maintains a focus on quality and safety.
Interestingly, Bard’s unveiling comes almost exactly two years after the publication of “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?” by former Google researchers, an event that led to the dismissal of Timnit Gebru, the former co-lead of Google’s ethical AI team.
The original research paper questioned “whether enough thought has been put into the potential risks associated with developing [large language models] and strategies to mitigate these risks.” Gebru’s public dismissal became a major news story.
Google stated in the announcement, “We’ll continue to improve Bard and add capabilities, including coding, more languages and multimodal experiences. And one thing is certain: We’ll learn alongside you as we go. With your feedback, Bard will keep getting better and better.”
To try Bard, users can sign up at bard.google.com.
