Proposed AI Transparency Act Introduced by U.S. Representatives

Written by Jeremy Werner

Jeremy is an experienced journalists, skilled communicator, and constant learner with a passion for storytelling and a track record of crafting compelling narratives. He has a diverse background in broadcast journalism, AI, public relations, data science, and social media management.
Posted on 01/09/2024
In News

In a push for greater transparency in AI, United States Representatives Anna Eshoo of California and Don Beyer of Virginia introduced the AI Foundation Model Transparency Act before the end of 2023. The Act targets AI foundation models, specifically those trained on broad data with over one billion parameters, applicable across various contexts like text, image, and video.

 

Under the proposed legislation, the Federal Trade Commission (FTC) would play a pivotal role in setting and establishing standards, requiring covered entities to disclose specific information about AI foundation models to the public. The Act seeks to address concerns related to copyright infringement and potential harms stemming from inaccurate or biased outputs.

 

The Act’s standards will delineate the information that must be submitted to the FTC and made publicly available. This includes details about training data sources, module documentation, data collection during use, and model limitations. Crucial aspects such as the source and composition of training data, labeling processes, model versioning, alignment with technical standards, benchmarks, limitations, and computational power used will be covered.

 

Enforcement of the regulations outlined in the Act falls under the jurisdiction of the FTC, with violations considered unfair or deceptive practices under the FTC Act. The Act’s requirements extend to common carriers, nonprofits, and entities over which the FTC has authority, provided they offer foundation models surpassing monthly use or output thresholds.

 

While the FTC retains the flexibility to adjust these thresholds as needed, a two-year timeframe has been set for the FTC to report to Congress and an annual update requirement for standards has been instituted. The introduction of the AI Foundation Model Transparency Act reflects a significant step toward establishing a framework for transparency and accountability in the rapidly evolving field of artificial intelligence.

If you’re curious how this potential bill, and other bills in the U.S. and around the globe could impact your company, don’t hesitate to contact BABL AI. One of their audit experts can offer valuable guidance and support.

Subscribe to our Newsletter

Keep up with the latest on BABL AI, AI Auditing and
AI Governance News by subscribing to our news letter