OpenAi, a foundation that deals with artificial intelligence, created in 2015 by Silicon Valley investors including Elon Musk, has made available a new chatbot model.
ChatGpt is able to understand human language and hold even very complex conversations.
In one week, it was tested by over 1 million users by visiting chat.openai.com.
ChatGpt is based on text samples taken from the internet (books, newspaper articles and web pages) to speed up obtaining answers on the most diverse topics.
A concern for giants like Google, should such an initiative one day turn into a full-fledged app, which would make web search obsolete.
However, ChatGpt is not without errors.
Stack Overflow, one of the most used sites by programmers, has banned the project,
"The main problem is that the indications produced by ChatGpt have a high error rate, even if they generally look good," says Stack Overflow in a post.
According to OpenAi, the project "may occasionally generate incorrect information and produce harmful instructions or distorted content. It also has limited knowledge of what happened after 2021."
To try ChatGpt, simply register an account on the OpenAi website or log in with a Google or Microsoft profile.
Just two of the giants who have tested artificial intelligence in open applications over the years, with little luck.
Microsoft bot Tay was retired in 2016 for a series of self-made racist remarks on Twitter while Google Lamda has been at the center of recent controversy,