My Blog
Technology

Google’s Bard can now read emails as company tries to show it’s useful

Google’s Bard can now read emails as company tries to show it’s useful
Google’s Bard can now read emails as company tries to show it’s useful


Google is integrating its Bard artificial intelligence chatbot with some of its other products, including Gmail, YouTube and Google Docs, as the tech giant tries to keep up with fierce competition and get more people to use its AI tools.

Bard, which competes with OpenAI’s ChatGPT and Microsoft’s Bing, will now be able to look through and summarize emails from Gmail, search through Google Docs and check flight prices with Google Flights, without users needing to leave the AI tool’s main screen. Bard appears as a box where a user enters a question. Until now, responses have been limited to simple text replies or photos from Google Images, but the latest updates mean YouTube videos, links to Google Doc files, and summaries of Gmail emails can appear embedded within Bard’s responses.

For example, a user could ask Bard to search through a tangled series of dozens of emails from a school and summarize only the most important items a parent needs to know.

The updates will give Bard access to its users’ content stored on Google products. The company does not use personal emails to train Bard and make it more competent, said Jack Krawczyk, the head of product for Bard. Google does, however, use Gmail emails to train its “Help Me Write” feature which provides suggestions for how to finish sentences and craft emails, unless users specifically turn it off.

The announcement Tuesday is the latest step in Google’s quest to push generative AI tools like Bard into more products and convince investors that it should still be considered the leader among AI companies.

It’s also a key test of whether customers of Google’s popular products will want to use — and eventually pay for — generative AI.

“There are many people who have found ways to use it but the average person is still saying ‘I haven’t found a way to integrate it into my personal life and I don’t necessarily fully trust these language models because they make things up,’” Krawczyk said.

AI chatbots still routinely spout false information a “hallucination” problem that researchers say is inherent in the underlying technology. Proponents of generative AI say the ability to make up information is important because it gives chatbots creative ability. Critics see it as a fundamental problem that calls into question whether the technology can be used for tasks that demand accuracy, such as work projects or education.

Google’s new approach to the issue is to run a separate Google search on the information Bard presents, looking for online articles that back up what it came up with on its own. Information that matches Bard’s suggestions gets highlighted in green, while sentences contradicted by other data sources get a dark orange highlight. Both provide links to the original web articles so users can check themselves.

Krawcyzk declined to say how many people regularly use Bard, which has so far been pushed out to millions of people and is available in most countries and in over 40 languages. There are some signs that people are beginning to lose interest in chatbots, such as the number of monthly users of ChatGPT dropping off earlier in the summer.

But companies are pouring huge amounts of money into developing the technology and running the computer servers necessary to keep it online.

“We still think it’s extremely early in this stage,” Krawcyzk. “People are still learning.”

Related posts

The Secret Ingredient in ‘The Great British Baking Show’

newsconquest

Google TV Streamer Voice Commands You Probably Don’t Know About

newsconquest

OpenAI Has ChatGPT and GPT-4 Updates Ready to Go. Here’s How to Watch on Monday

newsconquest