What Changed
SpaceXAI has filed a lawsuit against California's AB 2013, claiming that mandatory training data disclosures violate the First Amendment by compelling speech. The district court denied their request for a preliminary injunction, and the case is now before the Ninth Circuit Court of Appeals. The outcome could decide if free speech protections override state authority to enforce AI transparency.
This case has implications beyond California. If the court applies strict scrutiny and sides with SpaceXAI, similar transparency laws in California's SB 53, Illinois' SB 315, and New York's RAISE Act could be at risk. Even with intermediate scrutiny, companies may have stronger grounds to challenge and delay disclosure requirements during litigation.
Key Findings
The legal standard is crucial. SpaceXAI seeks strict scrutiny, which treats free speech as nearly absolute. California argues for rational basis review, where the state only needs a legitimate interest. The district court applied intermediate scrutiny, requiring California to show a valid consumer protection interest. The Ninth Circuit's choice of test will determine the fate of transparency mandates.
Trade secret and vagueness arguments are specific threats. SpaceXAI argues that AB 2013 forces the disclosure of trade secrets and is unconstitutionally vague. Success in these arguments could invalidate AB 2013 without affecting other laws. However, the First Amendment claim could set a precedent impacting any disclosure requirement.
Preliminary injunctions create uncertainty. Seeking preliminary injunctions can pause enforcement during litigation. Ben Rashkovich from Legal Advocates for Safe Science and Technology (LASST) notes that this practice can make AI regulations unpredictable. Your team must navigate a regulatory environment that shifts with litigation timing.
Industry opposition indicates broader resistance. An amicus brief opposing SpaceXAI's suit has nearly 30 co-signatories, including Americans for Responsible Innovation and the Electronic Privacy Information Center. This suggests the case tests whether First Amendment challenges can block AI governance measures.
What This Means for Your Team
Your compliance strategy now includes constitutional risk. If you're building governance controls around state transparency requirements, prepare for scenarios where these requirements are stayed or invalidated. This doesn't mean abandoning transparency efforts but structuring them to provide value beyond regulatory mandates.
Consider how your Technical Documentation (Annex IV) under the EU AI Act or your model validation evidence under SR 11-7 can serve dual purposes. Even if state laws become unenforceable, you'll need audit-ready documentation for federal exams, contracts, and voluntary frameworks like ISO/IEC 42001. Build your transparency infrastructure to satisfy multiple stakeholders.
You're also facing a coordination challenge. If preliminary injunctions create unpredictable enforcement windows, you can't rely on clear compliance deadlines. Your governance calendar needs flexibility to adapt to rapid changes in which requirements are enforced or stayed pending appeal.
Action Items by Priority
Immediate: Map your transparency obligations to legal stability. Inventory every disclosure requirement affecting your AI systems. Tag each by jurisdiction and legal basis. Identify which requirements are facing First Amendment challenges versus those grounded in federal guidance, international standards, or contracts. This mapping helps you see which controls might become optional and which remain mandatory.
This quarter: Decouple transparency from compliance minimums. If your only reason for maintaining training data documentation is "AB 2013 requires it," you're at risk. Reframe transparency work around operational needs: model debugging, vendor due diligence, and incident response. Build the business case so your documentation remains valuable even if the legal mandate disappears.
Next planning cycle: Prepare for injunction scenarios. Work with your legal team to define trigger events and response protocols. If a preliminary injunction pauses enforcement of a statute you're building toward, decide whether to continue implementation or reallocate resources. If an injunction lifts suddenly, can you adjust your timeline? Document these decision points now.
Ongoing: Track appellate arguments, not just outcomes. The Ninth Circuit's reasoning is more important than the win/loss result. If the court rules for California but applies intermediate scrutiny, companies can still challenge disclosure laws by disputing the state's interest. Subscribe to case updates and brief your governance committee on the legal standard adopted.
Strategic: Build transparency redundancy. Don't rely on a single legal basis for critical governance controls. If training data transparency is important for your risk management, anchor it in multiple frameworks: contractual requirements, ISO/IEC 42001's AI Management System controls, NIST AI RMF transparency objectives, and your own risk criteria. If one legal basis falls away, others keep the control in place.
