According to Beating AI News, U.S. Commerce Secretary Rutnick stated that artificial intelligence company Anthropic has repaired its relationship with the Trump administration, and the sharp differences between the two sides over national security issues that have lasted for months appear to have been resolved.

Lutnick stated that the two sides had previously had a "high-profile falling out," but Anthropic "has come to terms with it." He indicated that Trump and other government officials helped bridge the differences and secured the company's support, adding, "We've sorted things out, and now everyone is on the same page."

Previously, the U.S. Department of Commerce had briefly imposed export controls on two of Anthropic's top AI models in June, but lifted the restrictions two weeks later after Anthropic took steps to address the government's national security concerns.

According to Beating AI, Meta released the Muse Spark 1.3 model update on Wednesday, claiming that the version has significantly improved performance in coding and intelligent agent tasks.

Meta AI's head, Alexandr Wang, stated that the new model is "competitive with cutting-edge models" and will pave the way for future personal AI agent products, helping users achieve AI assistants that can work on their behalf around the clock. The price of Muse Spark 1.3 remains consistent with the previous version, a pricing strategy Wang described as "aggressive."

Meta also stated that its "contributor tier" option has been well-received by developers. This program significantly reduces the cost of coding products by allowing Meta to improve models using developers' work, and a "meaning double-digit percentage" of developers have already chosen this option.

Alexander Wang stated that as model capabilities improve, security has become a critical issue within Meta, and the company is increasing its investment in security and alignment. Muse Spark 1.3 will be available on Muse Code and the Meta API on the same day, with the highest inference version to be released after additional security testing.

According to a report by the WSJ, the U.S. government has filed documents with the court supporting OpenAI's position in its copyright dispute with publishers such as The New York Times, arguing that using news content to train artificial intelligence models complies with the "fair use" principle under U.S. copyright law.

The U.S. Department of Justice stated in a document that restricting AI companies from using copyrighted content to train large language models would violate fundamental principles of copyright law and hinder innovation and scientific progress. The New York Times sued Microsoft and OpenAI in December 2023, accusing ChatGPT and Microsoft Copilot of using a large amount of New York Times content to train models and generate responses.

The U.S. government argues that AI training is "highly transformative" to original works and will not harm the market value of copyrighted works; however, tightening fair use rules could result in only large technology companies being able to afford licensing fees, weakening the competitiveness of the U.S. AI industry. The parties involved are expected to file new motions after completing the exchange of evidence, requesting the court to rule on some of the core disputes.