Google is developing an artificial intelligence capable of writing news articles. The firm imagines it as a personal assistant for journalists.
The model has so far only been presented to American media, namely the washington post, NewsCorp and the New York Times.
Google does not want to replace journalists
Called Genesis internally, the tool is not to be confused with Google Bard, the Mountain View company’s ChatGPT competitor. This new technology is able to generate articles when it is provided with the information necessary to write them. The information was corroborated by a Google spokeswoman, who confirmed that ” in partnership with news publishers, especially smaller ones, we are exploring early ideas to potentially provide AI-enabled tools to help their journalists in their work “.
The model aims to automate certain tasks in order to free up time for others, and to help journalists in their choice of titles and writing styles. “Quite simply, these tools are not intended to, and cannot, replace the essential role of journalists in reporting, creating and fact-checking their stories. “said the spokesperson.
Google intends to develop responsible technology that can steer the publishing industry away from the pitfalls of generative AI.
Is it really a good idea?
On paper, the idea of an assistant AI for journalists may sound appealing, but will Google really be able to rid the model of language model bias? Generating AIs indeed tend to “hallucinate”, that is to say, they present false facts as if they were true. Moreover, the media CNET published many articles written by an AI and realized that the majority of them were wrong.
A study further found that articles generated by AIs spread false information. It is also possible that some publishers see such a tool as an effective way to save money, and therefore as a potential replacement solution. An OECD study recently concluded that artificial intelligence would jeopardize more than 25% of existing jobs in all rich countries.
Source : The New York Times
6