Sponsors (3)
The bill aims to reduce CSAM in AI datasets and clarify safety standards (helping children, developers, and investigators) but relies on a voluntary framework plus broad immunity that may limit immediate effectiveness, raise privacy risks, and reduce accountability while imposing costs on smaller actors.
Children and youth: the bill promotes detection, removal, and faster reporting of child sexual abuse material (CSAM) from AI training datasets, reducing exposure and aiding victim recovery.
AI developers, data collectors, and regulators: the bill establishes clear statutory definitions and names a Director (NIST role) to lead standards and guidance, reducing regulatory uncertainty about expectations.
AI developers and data collectors: the bill promotes standardized best practices and a voluntary framework for identifying and handling CSAM, which can reduce legal and reputational risk for organizations that adopt them.
Children and the public: the bill's voluntary framework and short timeline may produce uneven adoption across industry and limited immediate reduction in CSAM in models already in use.
Technology workers, nonprofits, and the public: broad civil immunity for actors who follow the Director's framework could reduce accountability and weaken incentives to minimize intrusive data collection or protect privacy.
Nonprofits, victims, and the public: increased scanning and reporting to law enforcement without specified safeguards raises privacy, surveillance, and false-positive concerns that could harm civil liberties and lead to improper investigations.
Based on analysis of 4 sections of legislative text.
Directs NIST to produce a voluntary framework to detect/remove/report child pornography in AI training datasets and grants limited civil immunity to compliant developers/data collectors.
Requires the Director of NIST to publish a voluntary framework (within one year) to detect, remove, and report child pornography in AI "covered datasets" used for training, developed with federal and private partners and public comment. Provides limited civil immunity for AI developers and data collectors who detect/remove/report child pornography in covered datasets in accordance with the framework, while excluding immunity for intentional misconduct, gross negligence, or violations of child-sex laws.
Official title: Require the Director of the National Institute of Standards and Technology to develop a framework for detecting, removing, and reporting child pornography in datasets used to train artificial intelligence systems, and for other purposes.
Introduced July 22, 2025 by John Cornyn · Last progress July 22, 2025