Google has announced the broader availability of its generative AI chatbot called Bard. The chatbot will now be accessible in over 40 languages and launched in the European Union (EU) after a delay due to data privacy concerns. Google has also introduced several new features to Bard, although some are initially available only in English.
Google first mentioned Bard in February as a response to the growing popularity of ChatGPT, an intelligent search engine/chatbot developed by OpenAI, which Microsoft supports.
In March, Bard was released in English for early access in the United States and the United Kingdom. The waitlist for Bard ended in May, resulting in a global rollout across 180 countries and support for Japanese and Korean languages. However, the EU launch was postponed due to concerns raised by the Irish Data Protection Commission (DPC), the data protection regulator for the EU. The DPC stated that Google had not provided enough information to address its data privacy concerns, even though it had informed the DPC about its plans to launch Bard in the EU.
With today’s launch, it appears that Google has addressed the DPC’s concerns and is making Bard available in the EU. Bard product lead Jack Krawczyk and VP of engineering Amarnag Subramanya claimed in a blog post, “We’ve proactively engaged with experts, policymakers and privacy regulators on this expansion.”
Alongside this expansion, several new features have been introduced to enhance Bard’s responses and increase its productivity potential. While some of these features were previewed in early May, they are now being broadly rolled out.
Enhancing User Experience with Customizable Tone and Style
One of the new features allows users to adjust the tone and style of Bard’s responses. There are five options available: “simple,” “long,” “short,” “professional,” or “casual.” Initially available only in English, users can toggle these options to align Bard’s default responses with their desired tone and style.
Furthermore, Bard can now vocalize its responses using a new text-to-speech AI feature. With more than 40 languages, users can access the chatbot’s audible reactions by clicking the new sound icon next to a prompt.
On the productivity front, Bard now offers enhanced code exporting capabilities, specifically allowing Python code to be shipped to Replit, the browser-based integrated development environment. Users can also include images with their prompts, enabling Bard to analyze the provided photo. To improve user experience, new options have been added, allowing users to pin, rename, and easily access recent conversations with Bard. Furthermore, Bard’s responses can now be easily shared with others through links.
According to Krawczyk and Subramanya, the creators of Bard, human curiosity and imagination are the driving forces behind creativity. Bard was designed to help users explore their curiosity, enhance their imagination, and ultimately bring their ideas to life. Bard’s purpose extends beyond answering questions; it aims to assist users in building upon their ideas.
Google has announced its most significant expansion, introducing its latest update to most parts of the world with added support for Arabic, Spanish, Chinese, German, and Hindi. Additionally, Bard, the chatbot, is now available in Brazil and in the EU.
Challenges in Bard’s Development and Google’s of Google
During the early stages oBard’s development, Google faced significant challenges in matching the quality of responses provided by competing chatbots like ChatGPT. Bard often gave factually incorrect answers accompanied by fabricated citations. This led to criticism from both Google employees, who referred to Bard as a” “worse than useless”s” and ” “pathological liar,” and a negative impact on Google’s stock, which briefly dropped by 8% aBard’s launch.
However, Google asserts that Bard has shown measurable improvements, particularly in mathematics and programming. Bard has also gained extensions, including support for Google’s apps and services and third-party partners like Adobe. Additionally, Bard can now explain code, present structured data in a table format, and incorporate images in its responses.
In another unfavorable incident for Google, Bloomberg recently reported that the human trainers responsible for teaching Bard often face overwork and inadequate compensation. Some contractors reportedly earn as little as $14 per hour, receive minimal training, and are expected to complete complex audits of Bard within minutes.
This revelation by Bloomberg follows an earlier report by Insider in April, which revealed that contractors assigned to test Bard were not given sufficient time to verify and cross-check Thchatbot’s most accurate answers. It appears that this situation has not improved.
