OpenAI and Microsoft allegedly knew AI training used copyrighted books

The documents also show that hiring of Tarun Gogineni in 2022 was meant to improve model writing quality with the expressed goal of “creating a machine that would supplant human authors.” Gogineni described authors’ complaints about stolen datasets as “acceptable economic disruption” and predicted “the death of the reader.” Microsoft’s awareness of OpenAI’s use of LibGen dates back to April 2019, and a 2022 memo from VP of Research Bob McGraw discussed removing LibGen files. The filings come as the case moves toward summary judgment.