As generative AI has gradually attracted attention from all walks of life, some controversial issues have also begun to surface. One of the biggest controversies is whether AI companies use copyrighted content in training models. Some writers, publishers, or media companies have also raised questions about whether AI companies use copyrighted content in training models.MicrosoftGenerative AI CompaniesA series of lawsuits were filed.
Mustafa Suleyman, the newly appointed CEO of Microsoft's AI division, spoke to CNBC on Thursday aboutGenerative AI gets data from the webHis reply was:
For content that is already publicly available online, the social contract for that content since the 1990s has been “fair use,” which means anyone canCopy, recreate, reproduceThese contents are generally considered "freeware".
There is also a separate situation where a website, publisher, or news organization explicitly states that its content may not be crawled or scraped for any reason other than to index it so that others can find it.Gray Area, I think it will gradually be resolved in the courts.
Image source: Pixabay
However, since last year, several news publishers or writers have sued Microsoft and OpenAI, accusing their chatbots of "stealing" their works to train AI models, infringing copyright. According to previous reports by IT Home, the institutions that have filed lawsuits against Microsoft and OpenAI include but are not limited to The New York Times and several Pulitzer Prize winners.