2024年2月15日 星期四

美國通 12 控告OpenAI 侵犯版權 (VS 公共利益i n the public interest.)的中心問題。 Angela Chao, C.E.O. of Family’s Shipping Company, Is Dead at 50


Only the authors’ central accusation against OpenAI remains. The group has claimed that the company “copied and ingested” their copyrighted work without permission or compensation. (One way that’s done is through so-called shadow libraries that house millions of texts online.) The end result, the authors argue: Every answer that ChatGPT creates is copyright infringement, because it was born from stolen work.


The economics of the generative A.I. industry are at stake in these cases. Rulings in favor of copyright owners could mean that companies like OpenAI would have to pay up to train their data-hungry systems. These tech companies argue that their programs are covered by fair use, and that their products’ output is sufficiently different from the original work — in short, that they aren’t violating copyright.

The outlook for such copyright battles remains unclear. While The Times’s lawsuit against OpenAI and Microsoft is still live, and recent rulings have been adverse to the authors group, a judge in a lawsuit involving A.I.-generated art generators said it was in the public interest.

  • In other A.I. news: OpenAI gave ChatGPT a better “memory” for user queries to improve future answers. And Andrej Karpathy, an A.I. researcher who co-founded OpenAI, has left the company.





Feb. 14, 2024

The Year Chatbots Were Tamed (美國各大AI 公司的聊天機械人如copilot, 過去一年將其大力整頓,變成無聊) , 各國駭客都用OpenAI Systems,

沒有留言: