Since launching its conversational AI Bard in February, Google has been improving the chatbot’s responses following its Twitter debut, including some misinformation. Recently, they have added generative AI features to nearly all their services apart from access to Bard, which was exclusive to a select few. Yesterday, Pixel users were invited to test out Google’s bot, and today the company announced that it is now available for anyone on Twitter.
The company spokesperson says:
“Starting to open access to Bard.””Bard did help us write,”
Sissie Hsiao and Eli Collins, Vice Presidents of Product and Research, respectively, encourage people to register at bard.google.com. The company has announced that they are starting to make the service available in the USA and UK today, with further expansion planned.
The company spokesperson went on to say:
“Expanding over time to more countries and languages.”
“The next critical step in improving it,”
The duo emphasized the importance of obtaining feedback from a larger group of testers.
You can communicate with Bard like you do with ChatGPT or Microsoft’s Bing AI – conversing using everyday language instead of a formal set of words. According to Hsiao and Collins, you could ask Bard for help achieving your goal of reading more books this year, get an easy-to-understand description of quantum physics, or seek out ideas for a blog post.
The duo remarked that Bard works as direct access to an LLM and is intended to be an additional option to Google Search. From the pictures shared in their announcement, it appears that the interface of Bard shares several similarities with Bing AI, but some vital aspects set them apart.
At the bottom of each response, four buttons are provided by Google: a thumbs up, a thumbs down, a refresh arrow, and one labeled “Google it.” Additionally, there is an option in the top right corner of the response called “View other drafts.” Bing AI does not provide these options, instead utilizing the space below each response for source citations.
Bard is driven by an agile and improved version of Google’s LaMDA, with the company announcing that it will be outfitted with more advanced versions in the foreseeable future. At its last I/O developer conference, Google demonstrated AI Test Kitchen using LaMDA 2, which they also revealed during that event. Meanwhile, Bing AI works on OpenAI’s GPT-4.
At the bottom of the screen, there are some differences between Bard and Bing. The former has a microphone icon at the end which implies speech-to-text is supported. Microsoft’s version has a broom logo on the left side of the input bar to help clear up any previous topics when starting something new.
It is worth pointing out that, beneath the text box of Bard, there is a line of small print stating, “Bard may show incorrect or offensive information which does not reflect Google’s opinions.”
In their statement, Google admitted that Large Language Models (LLMs) such as LaMDA are not flawless and can make errors.
Hsiao and Collins went on to say:
“As they are taught from a vast array of information which mirrors real-world biases and stereotypes, those may sometimes appear in their results.”
An instance of an error made by Bard was cited. It was stated that when inquired about a few recommendations for houseplants, Bard gave good advice; however, the bot erroneously identified the scientific name for the ZZ plant.
Google emphasized the importance of awareness of such difficulties and highlighted that considering quality and safety are major factors.
Hsiao and Collins say:
“We’ve also built in guardrails, like capping the number of exchanges in a dialogue, to try to keep interactions helpful and on topic,”
It is uncertain what the boundary for conversational exchanges in a dialogue may be, so we have requested additional details from Google. We will update this post when a response has been received.
Despite its potential drawbacks and possible mistakes, Google is convinced that investing in Bard will pay off, saying they intend to continue doing so.
Hsiao and Collins add to say:
“Continue to improve Bard and add capabilities, including coding, more languages and multimodal experiences.”
“It didn’t always get things right,”
“But even then, it made us laugh.”
Google’s move to open access to Bard AI chatbot represents a significant step forward in the evolution of AI-powered conversational interfaces. As technology continues to advance, Bard’s capabilities have the potential to transform various industries and create new possibilities for enhanced user experiences. Nevertheless, it is imperative to approach AI cautiously, considering ethical considerations and striving for responsible and ethical AI practices. With careful and responsible use, Bard and similar AI technologies can revolutionize how we interact with technology and shape the future of human-machine interactions.
Source: The Madison Leader Gazette