Alphabet‘s Google mentioned on Tuesday that Bard, its generative synthetic intelligence, may have the power to fact-check its solutions and analyze customers’ private Google information because the tech big scrambles to catch as much as ChatGPT in reputation.
The discharge final yr of ChatGPT, a chatbot from Microsoft-backed OpenAI, sparked a race within the tech business to present customers entry to generative AI know-how. On the time, ChatGPT was the fastest-growing client software ever and is now one of many high 30 web sites on the earth.
Bard has not taken off in the identical means. In August, it acquired 183 million visits, 13 p.c of what ChatGPT acquired, in line with web site analytics agency Similarweb.
Because it seeks to achieve floor within the fast-moving AI area, Google is rolling out Bard Extensions, enabling customers to import their information from different Google merchandise. As an example, customers may ask Bard to go looking their information in Google Drive or present a abstract of the person’s Gmail inbox. For now, Bard customers will solely be capable to pull info in from Google apps, however Google is working with exterior firms to attach their functions into Bard sooner or later, Google senior product director Jack Krawczyk mentioned.
One other new characteristic in Bard seeks to alleviate a nagging downside for generative AI: inaccurate responses often known as “hallucinations”. Bard customers will be capable to see which elements of Bard’s solutions differ from and agree with Google search outcomes.
“We’re presenting (Bard) in a means that it admits when it is not assured,” Krawczyk mentioned, explaining that the intention is to construct customers’ belief in generative AI via holding Bard accountable.
A 3rd new characteristic permits customers to ask others into Bard conversations.
© Thomson Reuters 2023
