Proposed Law Aims to Bring Transparency to AI Companies Through Copyright Disclosure

A new law, called the Generative AI Disclosure Act, has been proposed by US Representative Adam Schiff, aiming to enhance transparency from AI companies by requiring them to disclose the copyrighted works used in training datasets for AI systems. The bill would be retroactive and apply to existing and future AI systems, requiring disclosure prior to their release and maintaining a publicly available online database. Violators would face a civil penalty, with support from industry figures, and copyright guidance regarding AI-generated material is anticipated from the US Copyright Office later this year. This significant legislation is seen as a crucial step in ensuring fair treatment for creators in the age of AI.

Key Takeaways

• US Representative Adam Schiff introduced a bill to make AI companies disclose copyrighted works in their datasets, ensuring transparency and fair compensation.
• The proposed law would require AI makers to list referenced works and provide a URL to the dataset, with violators risking a civil penalty.
• The bill aims to safeguard the rights of creators and protect them from unauthorized use of their copyrighted materials in training generative AI systems.
• The US Copyright Office will issue guidance on deepfakes and AI outputs, but their plans for addressing AI training and licensing considerations remain unclear.
• Court decisions on significant copyright questions related to AI are anticipated to occur after the proposed Copyright Office guidance.

News Content

A proposed law introduced by US Representative Adam Schiff aims to bring more transparency to AI companies by requiring them to disclose the copyrighted works included in datasets used to train AI systems. The Generative AI Disclosure Act would retroactively apply to existing AI systems and would require creators to list referenced works and provide a URL to the dataset prior to public release. The bill has been championed as a crucial step in safeguarding the rights of creators and addressing the unauthorized use of copyrighted materials in training AI systems.

In related news, the US Copyright Office plans to issue guidance on deepfakes and AI-generated material, but the timeline for addressing the use of copyrighted works in training AI models remains unclear. The potential impact of this guidance on court opinions is yet to be determined, with some ongoing copyright lawsuits against AI companies not expected to be resolved until 2025. The proposed legislation and pending guidance reflect the ongoing efforts to address copyright issues in the rapidly evolving landscape of AI technology.

Analysis

The proposed Generative AI Disclosure Act by US Representative Adam Schiff seeks to enhance transparency in AI companies by mandating disclosure of copyrighted works in training datasets. This could impact AI firms like Google, Microsoft, and IBM, requiring them to list copyrighted works and provide dataset URLs. Short-term consequences may involve increased compliance costs and potential legal challenges. Long-term, the act may shape AI industry standards, protect creators' rights, and mitigate unauthorized use of copyrighted materials. The US Copyright Office's guidance on deepfakes and pending legislation reflects a broader effort to address copyright concerns, signaling a trend towards heightened regulation and accountability in AI development.

Did You Know?

• Generative AI Disclosure Act: Proposed legislation that aims to increase transparency in the use of copyrighted materials in training AI systems. The act would require AI companies to disclose the copyrighted works included in their datasets and provide a URL to the dataset prior to public release.
• Deepfakes and AI-generated material: The US Copyright Office plans to issue guidance on the use of deepfakes and AI-generated material, but the timeline for addressing the use of copyrighted works in training AI models remains uncertain. This reflects the ongoing efforts to address copyright issues in the rapidly evolving landscape of AI technology.