Shulou(Shulou.com)11/24 Report--

CTOnews.com, March 22 / PRNewswire-FirstCall-Asianet /-- Google just opened its chat robot Bard, a lightweight optimized version of the Google LaMDA language pre-training model, to British and American users yesterday, but the robot made an embarrassing mistake on the first day of public testing. When users asked how long it would be shut down by Google, it mistakenly replied that it had been shut down by Google.

It's not new for chatbots to have factual errors, which is one of the main problems with such big language models, and Google said bluntly yesterday that Bard's answer would contain factual errors, so why report the news? It is not intended to find fault with Bard, but this mistake is more interesting.

It turned out that the source of Bard's wrong answer was a joking post on Hacker News that Google would shut down Bard within a year. Google did shut down many of its own services in the past. Bard didn't understand the context and sense of humor of the post and shared it with users as a fact.

Google also admits that Bard contains some factual errors because it gets information from the Internet, and Microsoft's new Bing has a similar problem. This problem is one of the biggest challenges facing artificial intelligence based on the large language model (LLMs). Over time, both Google and Microsoft will improve the process of obtaining information, but occasionally there will be some problems. In addition, Bard is still in the preview stage, so it needs to go through more public testing.

However, CTOnews.com believes that how to make these chatbots better identify false news is a problem that these tech giants need to seriously consider.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.