Google’s artificial intelligence (A.I.) chatbot, Bard, was launched in March as an answer to OpenAI’s popular ChatGPT. However, Bard was initially deemed less capable and less conversational than its competitor. Google quickly revamped the tool with upgraded technology, but ChatGPT continued to dominate the chatbot landscape.
To surpass ChatGPT, Google recently announced its plan to integrate Bard with its most popular consumer services, such as Gmail, Docs, and YouTube. By doing so, Google aims to connect Bard with its vast range of online products.
While Bard has not received as much attention as ChatGPT, it has become a close contender. According to data from Similarweb, in August, ChatGPT had nearly 1.5 billion desktop and mobile web visits, three times as much as Google’s A.I. tool and other competitors.
Google’s product lead for Bard, Jack Krawczyk, acknowledged the limitations of the chatbot. Users have mentioned that Bard does not integrate well with their personal lives. Understanding these issues, Google released Bard Extensions, which provide capabilities similar to ChatGPT’s plug-ins.
With the latest updates, Bard aims to replicate some of Google’s search engine capabilities. It will include features such as Flights, Hotels, and Maps, allowing users to research travel and obtain transportation information. Furthermore, Bard is evolving into a personalized assistant that can inform users about missed emails and highlight important points in a document.
However, reliability remains a concern when it comes to A.I. chatbots. They are known to provide both correct information and falsehoods, leading to user confusion. To address this, Google has revamped the “Google It” button on Bard’s website. The button now double-checks Bard’s answers. When Google can support a claim with evidence, it will highlight the text in green and provide a link to a webpage backing up the information. When there is no factual basis, the text will be highlighted in orange.
In an attempt to build trust, Google is committed to acknowledging and rectifying its mistakes. However, the development of large language models like Bard has raised concerns about the use of consumers’ data. Google assures users that their personal information will be protected and not used for purposes such as ad targeting or the training of the Bard model.
Additional updates to Bard include enhancements to its underlying A.I., Pathways Language Model 2, which now accommodates image uploads in over 40 languages. Google also allows users to share Bard conversations with others, enabling them to see the responses and ask the chatbot further questions on the topic.
Despite Bard’s availability in over 200 countries and territories, Google still labels the tool an “experiment” rather than a finalized product. According to Krawczyk, it is the early days of this technology, and users need to have a clear understanding of its capabilities.
Overall, Google’s efforts to improve Bard’s functionality and reliability, along with its integration into popular consumer services, show a commitment to making A.I. chatbots more useful and trustworthy. As the technology continues to develop, it is crucial for companies like Google to prioritize user understanding and privacy.