Bipartisan Legislation Aims to Protect Personal Data in AI Development
Officers Josh Hawley (R-Mo.) and Richard Blumenthal (D-Conn.) have introduced bipartisan legislation designed to simplify the process for individuals suing tech companies. This move targets the use of unauthorized data in training artificial intelligence models, which they describe as “the biggest intellectual property theft in American history.”
The proposed AI Accountability and Personal Data Protection Act seeks to prohibit AI companies from using personal data or copyrighted works without obtaining explicit consent. This follows a recent congressional hearing where lawmakers criticized companies like Meta and OpenAI for their practices involving protected material.
“AI companies are blinding Americans while disregarding the rights of artists, writers, and creators,” Hawley asserted in a statement. “It’s time for Congress to ensure that American workers can protect their personal data and creative output.”
This bill would allow individuals to sue if their personal data or copyrighted works are used without their permission. Even with consent, companies would be required to inform users about any third parties accessing the data, along with the imposition of financial penalties and injunctive relief.
Meta and OpenAI have been approached for comments regarding this legislation.
Hawley further stated, “This bipartisan law ultimately empowers everyday Americans who are trying to navigate the chaotic landscape of Big Tech.”
Blumenthal, who is a Democratic co-sponsor of the bill, emphasized the serious risks to privacy that warrant such legal measures. “Technology companies must be held accountable for any unauthorized collection, monetization, or sharing of personal information,” he remarked.
In recent years, various high-tech firms have faced lawsuits from content creators and publishers claiming that their copyrighted works were “scraped” for AI model training.
For instance, Thomson Reuters successfully appealed against Ross Intelligence for using Westlaw’s copyrighted legal materials to develop legal research AI. A federal court ruled in February that Ross had, in fact, committed copyright infringement.
The news agency is currently pursuing unspecified damages.
In December 2023, the New York Times filed a lawsuit against OpenAI and Microsoft, alleging that their articles were used to train systems like GPT-4 without authorization. That case is still pending.
Last month, a federal judge noted that using books to train AI models could be considered “transformative” and, therefore, fair use. However, keeping a direct copy—the “pirated” version—poses issues of direct infringement. The question of damages and remedies remains unresolved.
Additionally, a case involving Meta’s use of books without permission for training language models is ongoing. The court has ruled that their usage was “very transformative” and qualifies as fair use, but the liability regarding stored “pirated” material remains to be determined.
