AI-Washing Under Scrutiny: Texas AG Investigates DeepSeek's Claims
March 6, 2025
Privacy Plus+
Privacy, Technology and Perspective
This week, let’s revisit exaggerated AI claims in light of recent state-level action against what some call "AI-washing." For more background, you can read our previous post in January examining the FTC's crackdown:
https://www.hoschmorris.com/privacy-plus-news/ai-marketing-claims-under-ftc-scrutiny
The DeepSeek Investigation
On February 14, 2025, the Texas Attorney General announced a comprehensive investigation into DeepSeek, the Chinese artificial intelligence company whose advanced technology recently made headlines in financial markets, in particular. The investigation highlights concerns about overstated AI capabilities and data privacy practices.
The Texas Attorney General's probe centers on multiple aspects of DeepSeek's operations:
· AI Capability Claims: DeepSeek claims that its AI model rivals the most advanced systems globally, including OpenAI's Model o1. The investigation is scrutinizing these claims for potential misrepresentation or unfair competition under Texas law.
· Privacy Practices: The Texas AG has notified DeepSeek that its platform allegedly violates the Texas Data Privacy and Security Act, which protects consumer data. This raises questions about whether companies making bold AI claims are equally attentive to their data protection obligations.
· Third-Party Distribution: The Texas AG has also issued Civil Investigative Demands (CIDs) to Google and Apple, requesting analyses of the DeepSeek application and submission documentation. This suggests that regulators are also examining how platforms that host AI applications verify the claims made by developers.
You can read the Texas AG's press release here:
National Security Dimension
Compounding the AI-washing concerns are DeepSeek's alleged ties to the Chinese Communist Party, and concerns that DeepSeek may:
· Serve as a proxy for collecting data from Texas residents
· Potentially compromise individual privacy
· Undermine American leadership in AI development
These concerns prompted Texas to ban the DeepSeek and other Chinese technologies from all government-issued devices in January 2025. You can read the Governor’s announcement of that ban by clicking on the following link:
https://gov.texas.gov/news/post/governor-abbott-announces-ban-on-chinese-ai-social-media-apps
Our Thoughts
The End of AI's "Magical Early Days": The DeepSeek investigation appears to be an extension of the trend we described in our previous post. For some time we have lived in AI's "magical early days," when merely invoking its presence (“our product has AI!”) seems to bestow credibility on products and services. But now, regulators are becoming increasingly focused on ensuring these claims are substantiated.
Truthful Advertising Requirements Across Jurisdictions: The Texas investigation highlights how the requirements for truthful advertising apply to AI claims at both federal and state levels. Just as with the FTC's Prior Substantiation Rule, we believe states will increasingly require that companies have in hand reasonable proof of their AI capabilities before making public claims about them.
Triple Threat: AI Claims, Privacy, and National Security: The Texas investigation adds another layer of complexity: the intersection of possibly exaggerated AI claims with privacy compliance and national security concerns. When foreign companies make bold assertions about AI capabilities while potentially mishandling consumer data, they will (and in our view, should) face a perfect storm of regulatory scrutiny.
Due Diligence: Verifying AI Capabilities: Recent regulatory actions highlight the need for rigorous due diligence when procuring AI systems. Rather than accepting marketing claims, businesses should verify AI capabilities through concrete evidence. Start by requesting model cards—standardized documentation that outlines an AI system's training data, performance metrics, limitations, and ethical considerations. These model cards provide the essential foundation for informed procurement decisions.
Contractual Safeguards for AI Procurement: When drafting contracts, explicitly reference the documented capabilities through specific representations, warranties, and indemnification clauses. These contractual protections should clearly address performance parameters and potential legal exposures.
Preparing for the New Regulatory Landscape: The increasing coordination between federal and state regulators marks a clear shift toward greater accountability in AI marketing. Companies that implement formal vendor assessment protocols—including third-party validation for high-risk AI applications—protect both their operations and their legal position as enforcement intensifies.
---
Hosch & Morris, PLLC is a boutique law firm dedicated to data privacy and protection, cybersecurity, the Internet and technology. Open the Future℠.