In today’s news and commentary, the United Steelworkers President urges President Trump to block a proposed steel merger; Goodwin Procter discloses diversity data to the EEOC; and a federal judge in California hints that authors’ work might be subject to fair use by Anthropic AI.

Yesterday, United Steelworkers President David McCall urged President Donald Trump to block Nippo Steel Corporation’s bid to buy United States Steel Corporation – a proposed merger that has garnered criticism for potentially consolidating the steel industry. McCall asserted that, “Allowing the sale of U.S. Steel to Nippon, a serial trade cheater, will be a disaster for American Steelworkers, our national security and the future of American manufacturing… President Trump has publicly pledged to block this sale since January 2024… We now urge him to act decisively, shutting the door once and for all on this corporate sellout of American Steelworkers and defending U.S. manufacturing.”

Goodwin Procter, one of the world’s largest law firms, recently sent the Equal Employment Opportunity Commission (EEOC) over 200 pages of information on diversity data that the agency had requested. Six other law firms that the EEOC requested information from – Kirkland & Ellis, Latham & Watkins, Simpson Thacher, A&O Shearman, Skadden, and Milbank – have already made deals with the Trump administration to provide pro bono legal services in exchange for being released from the investigation. The information Goodwin disclosed to the EEOC revealed that the firm ended its participation in the Sponsors for Educational Opportunities legal fellowship, which offers underrepresented law students the opportunity to work for prestigious law firms. Additionally, Goodwin halted its relationships with the Leadership Council for Legal Diversity and the Mansfield certification program – a credential that indicates the degree of a law firm’s efforts to increase diversity, equity, and inclusion internally.

In a hearing yesterday of the case Bartz v. Anthropic, Senior Judge of the U.S. District Court for the Northern District of California William Alsup suggested he was “inclined” to find that Anthropic violated the Copyright Act when the company made initial copies of pirated books to train its artificial intelligence (AI) model, but that subsequent uses were subject to “fair use.” Anthropic argued that the four factor fair use test points in its favor. Three weeks prior to this hearing, Judge Vince Chabria, also of the U.S. District Court for the Northern District of California, heard a similar case in which several authors brought suit against Meta for allegedly engaging in copyright infringement for training its latest AI model “Llama 3” on the authors’ published works. Judge Chabria suggested that Meta’s AI model could destroy markets for copyrighted work, but that the evidence offered by the authors asserting harmful market effects was lacking. The decisions of these cases will have lasting effects on the artists, writers, and other creatives who have produced, and will continue to produce, work that AI companies use without permission to train their AI models.