Senate Hearing Examines AI Industry's Use of Copyrighted Works for Training

Strategic Insights -

A recent Senate Judiciary Subcommittee hearing highlighted the complex relationship between AI companies and intellectual property rights, particularly copyrights. The hearing, entitled "Too Big to Prosecute?: Examining the AI Industry's Mass Ingestion of Copyrighted Works for AI Training," brought together experts to discuss the impact of AI on copyrighted works and the industry's reliance on these works for training. Senator Richard J. Durbin, the chair of the subcommittee, questioned witnesses about the fairness of allowing AI companies to use copyrighted works without permission, arguing that this practice could have significant economic consequences for authors and creators.

Key Takeaways:

  • Senator Durbin questioned witnesses about the use of copyrighted works for AI training, highlighting the potential economic consequences for authors and creators.
  • Edward Lee, a law professor at Santa Clara University, stated that the initial burden of "fair use" is on the defendant, and that AI companies are ultimately the winners in the approach being taken.
  • Maxwell Pritt, a partner at Boies Schiller Flexner LLP, representing plaintiffs in Kadrey v. Meta Platforms, testified that Meta did not compensate copyright owners for the use of their works in training its GenAI model.
  • Senator Durbin asked Prof. Lee if authors should be prepared to pay the price for AI companies to use their copyrighted works without permission, to which Prof. Lee responded that the established Supreme Court precedent on "fair use" should be protected.
  • Meta spent money on contributing processing power to pirate from illicit websites and paid Amazon to host pirated data, according to Mr. Pritt.
  • The downloading and uploading of pirated copyrighted material could meet the mens rea requirement of willfulness necessary for criminal infringement, according to Mr. Pritt.

Statistics:

  • Meta spent money on contributing processing power to pirate from illicit websites and paid Amazon to host pirated data (Source)
  • The United States has a priority in AI development, with winning the AI race being important in an arms race with China (Source)
  • In the case of Kadrey v. Meta Platforms, authors, including Richard Kadrey and Sarah Silverman, sued Meta alleging copyright infringement related to the training of Meta's LLaMA AI model using copyrighted books (Source)

Sources:

  • Durbin, Richard J. "Too Big to Prosecute?: Examining the AI Industry's Mass Ingestion of Copyrighted Works for AI Training." Senate Judiciary Committee, 2023.
  • Lee, Edward. Testimony before the Senate Judiciary Subcommittee. 2023.
  • Pritt, Maxwell. Testimony before the Senate Judiciary Subcommittee. 2023.
  • "Kadrey v. Meta Platforms, Inc." Boies Schiller Flexner LLP, 2023.