外刊：The NYT sues MS and OpenAI for using its texts without permission

The New York Times sues Microsoft and OpenAI for using its texts without permission

Oliver Thansan / USANews

26 December 2023 Tuesday 21:27

The New York Times filed a lawsuit this Wednesday against Microsoft and the artificial intelligence company OpenAI for using their texts without permission to train their AI models.

According to the media, millions of its articles have been used to fine-tune chatbots against which they are now forced to compete in the information sector.

"Through Microsoft's Bing Chat (recently renamed Copilot) and Open AI's ChatGPT, the defendants seek to take advantage of The Times' enormous investment in its journalism, using it to build substitute products without permission or payment," the statement reads. lawsuit, filed in a Manhattan court.

The newspaper is not seeking specific financial compensation, but is seeking to hold the defendants liable for "billions of dollars" in damages, and to destroy AI models that use information copyrighted by The New York Times.

As part of the lawsuit, examples of various texts produced by GPT-4 (one of OpenAI's products) are included that are almost indistinguishable from some research published by the medium.

Additionally, they show that Microsoft's Bing search engine can be asked to copy entire paragraphs of news from the Times, which requires a subscription to read much of its content.

Artificial intelligence chatbots like ChatGPT use huge amounts of text data to predict the most likely word in response to a question, recreating human speech with astonishing accuracy.

However, in many cases, all those texts that are used to train the model, such as books or press articles, are protected by copyright, and more and more authors and companies demand to be compensated for the use of their work.

At the beginning of the month, OpenAI, whose main investor is Microsoft, reached an agreement with the company Axel Springer, which publishes the media Politico, Business Insider or Bild, to use its content in exchange for a fee.

The New York Times sues Microsoft and OpenAI for using its texts without permission

《纽约时报》起诉微软和OpenAI未经许可使用其文章内容

The New York Times filed a lawsuit this Wednesday against Microsoft and the artificial intelligence company OpenAI for using their texts without permission to train their AI models.

《纽约时报》本周三对微软和人工智能公司OpenAI提起诉讼，指控其未经许可使用他们的文章内容来训练AI模型。

According to the media, millions of its articles have been used to fine-tune chatbots against which they are now forced to compete in the information sector.

该媒体称，其数百万的文章被用于对聊天机器人进行微调，现在他们被迫在信息领域与之竞争。

"Through Microsoft's Bing Chat (recently renamed Copilot) and Open AI's ChatGPT, the defendants seek to take advantage of The Times' enormous investment in its journalism, using it to build substitute products without permission or payment," the statement reads. lawsuit, filed in a Manhattan court.

“微软的Bing Chat（最近更名为Copilot）和Open AI的ChatGPT，被告试图利用《纽约时报》对其新闻业的巨大投资，在未经许可或付款的情况下使用其制造替代产品”，在曼哈顿法院提起的诉讼的声明中提到。

The newspaper is not seeking specific financial compensation, but is seeking to hold the defendants liable for "billions of dollars" in damages, and to destroy AI models that use information copyrighted by The New York Times.

该报不是要求具体的经济赔偿，而是要求被告承担“数十亿美元”的损害责任，并销毁使用《纽约时报》版权信息的AI模型。

As part of the lawsuit, examples of various texts produced by GPT-4 (one of OpenAI's products) are included that are almost indistinguishable from some research published by the medium.

GPT-4（OpenAI的产品之一）产生的各种文本的例子被列为了诉讼的一部分，这些文本和该媒体发布的一些研究几乎无法区分。

Additionally, they show that Microsoft's Bing search engine can be asked to copy entire paragraphs of news from the Times, which requires a subscription to read much of its content.

此外，他们还展示了微软的Bing搜索引擎可以被要求拷贝《纽约时报》的整段新闻，这些新闻是需要订阅后才能阅读其大部分的。

Artificial intelligence chatbots like ChatGPT use huge amounts of text data to predict the most likely word in response to a question, recreating human speech with astonishing accuracy.

像ChatGPT这类人工智能聊天机器人会使用大量的文本数据来预测一个问题最有可能回答的单词，以惊人的准确度来重现人类说话。

However, in many cases, all those texts that are used to train the model, such as books or press articles, are protected by copyright, and more and more authors and companies demand to be compensated for the use of their work.

然而，很多情况下，所有这些用于训练这些模型的文本，比如书本或者新闻文章，是受版权保护的，而且越来越多的作者和公司要求对使用他们的作品进行补偿，

At the beginning of the month, OpenAI, whose main investor is Microsoft, reached an agreement with the company Axel Springer, which publishes the media Politico, Business Insider or Bild, to use its content in exchange for a fee.

本月初，主要投资者是微软的OpenAI与发布Politico、Business Insider和Bild的Axel Springer公司达成协议，付费使用其内容。