Elon Musk’s artificial intelligence company, xAI, has filed a lawsuit challenging California Assembly Bill 2013 (“AB 2013”), a legislation mandating generative AI developers to disclose detailed information about the datasets they used to train their models.
AB 2013, also known as the Generative Artificial Intelligence: Training Data Transparency Act, was signed into law on September 28, 2024, and was set to take effect on January 1, 2026. The lawsuit, filed in federal court in the Central District of California at the tail end of 2025, seeks to permanently enjoin California AG Rob Bonta from enforcing the provisions of AB 2013 against xAI.
xAI’s complaint emphasizes the substantial resources it dedicated to identifying high-quality data to train its flagship AI chatbot, Grok, and its efforts to safeguard information about the datasets it uses, arguing that the datasets themselves and their usage are valuable trade secrets. According to xAI, “AB 2013 is [] a trade-secrets-destroying disclosure regime that hands competitors a roadmap to learn how companies like xAI are developing and training their proprietary AI models.” xAI also argues that the information sought to be disclosed by AB 2013 is not valuable to consumers, “who are far more interested in evaluating how an end-product performs the tasks it is given than in obtaining technical details about the datasets and processes companies use to train their models.”
The lawsuit alleges that AB 2013 runs afoul of the U.S. Constitution’s Takings Clause by compelling xAI to give up its trade secrets – or significantly reducing their value – without any promise of compensation. xAI also advances a First Amendment argument, stating that forcing it to disseminate specific information unlawfully compels speech, which is treated no differently than restricting speech. Looking to AB 2013’s legislative history, xAI notes that the law was designed to “help[] identify and mitigate biases,” arguing that it therefore triggers the strict scrutiny standard required by content- and viewpoint-based regulation of speech.
With no dearth of litigation funds on either side – xAI is valued at well over $200 billion – this lawsuit looks set to go the distance. Interestingly, the filing comes amidst international backlash over Grok reportedly generating sexually explicit content featuring minors. A 2023 study conducted by Stanford University had found thousands of child sexual abuse images in a dataset used to train AI image generation models.
CIPAWorld 