A real tidal wave. Currently in testing, ChatGPT, the artificial intelligence-based chatbot developed by OpenAI, has gone viral, surpassing 1 million users since it went public last week, he said in a tweet. Posted this Monday. , December 5, the CEO of the company Sam Altman. A dazzling success that can be explained by the impressive writing capabilities of this tool, as well as its ease of use.
In fact, ChatGPT, created by the Californian startup behind DALL-E, is a prototype conversational robot powered by artificial intelligence capable of understanding human language, producing highly accurate written text, and simulating conversations in an extremely believable way. “The dialog format allows ChatGPT to answer questions, acknowledge errors, or challenge and reject inappropriate requests,” says OpenAI. in a press release released last week.
To achieve this, this technology has been tested on a large amount of text: newspaper articles, novels, film and television scripts, as well as online conversations. Learning that allows you to understand the context of a conversation and give relevant and coherent answers.
A future competitor to Google?
According to Sam Altman, ChatGTP is only in its first iteration. If its development continues to accelerate, the tool could be, in the medium term, a credible alternative to Google due to its ability to formulate complex responses to Internet users’ requests, underlines The Guardian.
In addition, ChatGPT, which can be adapted to a wide range of topics and language styles, could also be used to generate content for websites, social networks, or write scripts for movies or series. A skill that questions the place of the human being that this technology could well replace in the long term.
However, a concern is tempered by the current technical limitations of ChatGPT: a lack of nuance, critical thinking, and the ability to make ethical decisions. Essential elements in the writing of a certain type of content, especially journalistic. In addition, the tool is not foolproof and can provide completely false answers by writing “answers that seem plausible but are incorrect or nonsense,” the company acknowledges. A problem that OpenAI recognizes as difficult to solve.
Source: BFM TV
