As significant consumer attention and excitement revolves around generative artificial intelligence (AI) and chatbots in particular, the pace of development shows no signs of slowing down. Google is now giving its chatbot – Bard – a set of significant and wide-ranging updates, including more Indian and worldwide language knowledge, responses that can be read out, saving a history of your conversations with Bard and the ability to export Python code generated using Bard (a capability that too was enabled recently).
In what is safe to file as Bard’s biggest update yet, milestones till now have included Bard’s official unveiling earlier this year, though the pivotal moment came in summer when Bard was made available across 180 countries to all users. It also coincided with the switch to using PaLM 2 large language model, or LLM.
From today, Bard will work with 40 more global languages, in addition to existing support for English, Japanese and Korean. The Indian languages will include Hindi, Tamil, Telugu, Bengali, Canada, Malayalam, Marathi, Gujarati and Urdu. Global language support now includes Arabic, Chinese, German and Spanish.
Bard also got rolled out to 59 new countries, including within the European Union (EU) and Brazil.
“It will be in line with our bold and responsible approach to AI. We’ve proactively worked with a number of policymakers, regulators and experts on this roll-out, and as we continue to evolve, we will keep our AI principles as a guide,” says Amar Subramanya, VP of Engineering at Google, in a briefing that HT was part of.
These updates come at a time when Google Bard is competing with OpenAI’s ChatGPT and Microsoft’s Bing chatbot, which has similar underpinnings to ChatGPT. Bard’s widening feature set gives it more tools in battle.
In a follow-up to the text-to-image feature that was already part of Bard, the reverse will be true now too. A user will be able to input an image for Bard to search with (Google Lens integration comes along too). Generative AI will attempt to generate details about what it decodes from the image you’ve shared with Bard. Think of this as Google Lens, but potentially much smarter.
“This past weekend, I had this chore of cleaning up one of the desks in our home and I ended up taking a picture if it and going to Bard. Hey, can you give me ideas on how I can declutter this desk, it gave me a bunch of suggestions and it was awesome to get help in a moment that mattered,” shares Subramanya.
Google is also enabling conversation history in Bard, something that was conspicuously missing thus far, that’ll allow users to revisit previous chats and responses. There will also be an option to choose the tone and style of responses you’d like from Bard – shorter, longer, professional or casual. For now, this vibrancy will be limited to English language chats, but expect wider support in the coming months.
Microsoft has similar response options for Bing too – more creative, more balanced, more precise, though the differences may not be as apparent in all questions a user may send its way.
Bard chatbot, which till now was primarily geared towards text responses, will now be able to speak out those responses too. “If you generate a poem, you might actually want to listen to what was generated rather than just read it,” says Subramanya. “It’s also sometimes super useful to hear a pronunciation,” he adds.
Recently, Google Bard adopted the ability to generate Python code for app and software developers. Now, Bard will also be able to export it to Replit, a tool that lets developers run code live within a web browser.
Google Bard, which is still very much in the experimental stage (and Google reemphasises that) is available on the web browser, but unlike Microsoft Bing and ChatGPT (in an early stage of rollout), does not have apps for Android phones or Apple iPhones for now. Bing, in fact, is also integrated within the Edge web browser apps and the SwiftKey keyboard app for smartphones.
At the Google I/O developer conference earlier this summer, Google hinted at focus on integrating Bard within Google’s wide tool set, including Google Lens, Mail and Search.
Microsoft, in March, had confirmed the Bing chatbot had clocked 100 million active users within weeks of launch for consumers. Google hasn’t released any official statistics on Bard’s active user base yet, but insists the more user feedback they get, the better they’ll be able to improve Bard.
Over the summer, Bing’s importance to develop the Microsoft 365 Copilot tool was highlighted at the company’s annual developer conference keynote. It’ll arrive with integration within Windows 11 and Microsoft 365 apps including the office suite, later this year.
Google is looking to plug in Bard into its own suite of apps, including Docs, Drive, Gmail and Maps. There will be third-party integrations as well, including with Adobe Firefly to help users generate high quality AI images.
“The more people use it and give us feedback, the more opportunity we have to make this model and generative AI better. We will be able to figure out how we can actually bring it to more users in a safe and responsible way,” says Subramanya.
The new features set for Bard rolls out for all users, though some functionality may arrive in batches.
- ABOUT THE AUTHOR
Vishal Mathur is Technology Editor for Hindustan Times. When not making sense of technology, he often searches for an elusive analog space in a digital world.