Google's major Bard expansion now scans images and talks to you
Google has released a what it calls its biggest expansion to date of its AI assistant, Bard, with new, useful features such as pinned conversations and more flexible output selections that include a “long mode” for more detailed replies.
The ongoing war between Bard, Microsoft’s Bing and OpenAI’s ChatGPT has expanded on several fronts. To date, Bing is one of the only chatbots that you’ve been able to interact with via a mic, using the mobile version of Bing. Google has taken the opposite approach: now Bard will talk to you, orally pronouncing words. Bard will also communicate in not just English, but also in Arabic, Chinese, German, Hindi and Spanish, and access to the Bard chatbot has opened up inside Europe, now, too.
For most users, the three most useful features arriving in the new update include the new pinned threads, the expanded conversational modes and image analysis. ChatGPT assigns a topic header to each new conversation (which Google calls threads), and you can skip back and forth between your conversational history to find the information you’re looking for. Now Bard will be able to do that, too, so you’ll be able to return to an earlier conversation. (A Google representative says that there’s no limit to how many threads you can pin, or how far back they’ll go.) While you can pin a conversation for your own benefit, you’ll also be able to share it with a friend, too.
Bing was also the first to launch various conversational modes by default: Balanced, Precise, and Creative. Google’s Bard is taking a slightly different approach: You’ll be able to choose simple, long, short, professional or casual instead. This feature is only live in English, though. One advantage of ChatGPT’s GPT-4 update is its support for lengthier inputs and outputs, and it appears Bard is trying to match ChatGPT’s capabilities here.
YouTube / Google
Finally, Google is adding the image analysis tools it announced at its Google I/O developer conference into Bard as well. With Bard, you’ll be able to analyze the contents of an image — what this will mean, in reality, is somewhat up in the air. In practice, that means that you’ll be able to upload an image, as you can do with Google Image Search, and ask question about it using Bard. However, it will not include image generation quite yet.
Google appears to be stepping through the features it announced at Google I/O, though they’re not all here quite yet. Bard’s projected ability to produce code and then potentially run it isn’t here, though it can be exported to a Google Collab test server as well as another third party, Replit.
Still, Bard is free, available, and Google seems just as dedicated to pushing it as mainstream as possible. You should be able to try out Bard’s new capabilities now.
Author: Mark Hachman, Senior Editor
As PCWorld’s senior editor, Mark focuses on Microsoft news and chip technology, among other beats. He has formerly written for PCMag, BYTE, Slashdot, eWEEK, and ReadWrite.
Recent stories by Mark Hachman:
Comcast’s new NOW prepaid Internet looks surprisingly compellingUdio’s AI music is my new obsessionBroadband ‘nutrition labels’ kick in, revealing hidden fees for ISPs