Generative AI is smart, but it’s not reliable. This is one of the major issues holding the technology back from being the major frontier that companies claim it to be. Google is working on making its chatbot Bard more reliable, and it’s going to launch a feature that will help in that regard. According to Axios, you’ll soon be able to ask Bard to double-check its answers.
Something that Google won’t live down soon is the incident during Bard’s unveiling. When asked to give an answer to a science question, Bard gave a bit of inaccurate information. This didn’t only cause the company’s stocks to tank, but it also caused us to question the reliability of generative AI as a whole.
You will soon be able to have Bard double-check its answers
This issue is still ongoing, but Google is working on solving it. The company will allow people to install the Bard Extensions plugin when they use the chatbot. This plugin will connect Bard to several of Google’s products. These include products like Maps, Docs, Gmail, YouTube, Drive, and Google Flights. This will add Bard’s generative AI abilities to these services.
You can use the plugin to generate all kinds of text on these platforms. You can use it to write up a draft for an email, summarize documents, etc. However, what if you have doubts about the content being produced? Google has something to help.
There will be a little Google button next to the responses that Bard generates. When you click on that option, Google will go through the answer, line by line, and double-check it. It will pull information from the internet to fact-check itself. So, if you have doubts about an answer, the company has a safety net for you.
This new functionality is rolling out to English-speaking users. We’re not 100% sure about the rate of the rollout. We were able to get the feature here at Android Headlines, so if you don’t see it, you’ll want to wait just a bit.