据The Information报道,人工智能研究公司OpenAI计划向媒体出版公司支付数百万美元的年费,以购买新闻许可,用于训练其大语言模型。消息透露,OpenAI已经向一些媒体公司开出了每年100万至500万美元的价格,以获得将新闻内容用于训练自家大语言模型的授权许可。

据悉,苹果公司也正在努力追赶OpenAI和谷歌在生成人工智能领域的领先地位,并试图与出版商就其内容的使用达成交易。有高管表示,苹果的出价甚至更高。

这一举动凸显了人工智能公司在获取高质量数据以训练其算法方面的重要性。新闻内容因其丰富的话题和多样的表述方式,成为训练自然语言处理模型的宝贵资源。

With its substantial investment in purchasing news content licenses, OpenAI aims to enhance its large language model training. According to a report by The Information, the artificial intelligence research company is planning to pay millions of dollars annually to media publishing companies for news licenses, which will be used to train its large language models. The report reveals that OpenAI has offered prices ranging from $1 million to $5 million per year to some media companies for the authorization to use their news content for training its own large language models.

It is understood that Apple is also striving to catch up with OpenAI and Google in the field of generative artificial intelligence and is trying to reach deals with publishers regarding the use of their content. Executives have stated that Apple’s offers are even higher.

This move highlights the importance for artificial intelligence companies to obtain high-quality data to train their algorithms. News content is considered a valuable resource for training natural language processing models due to its rich topics and diverse expression styles.

【来源】https://www.theinformation.com/articles/openai-offers-publishers-as-little-as-1-million-a-year

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注