Blackburn Rolls Out 300-Page AI Plan to Create One Federal Rulebook
The Senate’s draft proposal for national artificial intelligence (AI) legislation, unveiled on Wednesday, aims to preempt state laws by focusing on regulations that protect the “4 Cs” – children, creators, conservatives, and communities.
Led by Sen. Marsha Blackburn of Tennessee, the measure, dubbed the TRUMP AMERICA AI Act, spans nearly 300 pages. This proposal follows a December executive order from former President Donald Trump, which aimed to create a unified AI framework to avoid a confusing patchwork of state regulations that could stifle innovation.
Blackburn’s proposal marks the first legislative effort since Trump’s order, formalizing two key federal resources related to AI:
- The Center for AI Standards and Innovation (CAISI) at the National Institute of Standards and Technology (NIST).
- A governance and funding structure for the National Science Foundation’s (NSF) National Artificial Intelligence Research Resource (NAIRR) pilot.
In a statement, Blackburn emphasized the need for Congress to establish a single federal rulebook for AI to ensure that America remains competitive in the global race for AI dominance.
Child Safety Regulations
The proposal includes measures to protect users under the age of 17 on online and social media platforms. Key requirements include:
- Implementing tools and guardrails designed to enhance user safety.
- Adopting safer design practices and adding privacy and parental control tools.
- Restricting research involving children and teens and improving transparency over algorithm-driven content.
Additionally, platforms must verify the ages of underage chatbot users through government-issued IDs and disclose that bots are not human or licensed professionals. Non-compliance could lead to penalties of up to $100,000.
A Kids Online Safety Council would also be established to advise Congress on emerging online risks to minors.
Protections for Creators
Under Blackburn’s draft proposal, copyright holders would gain a legal tool requiring transparency about how AI models are trained. Creators suspecting their work has been used to train generative AI systems could request a court-issued subpoena to disclose the training data.
The proposal mandates federal agencies, particularly NIST, to create standards for identifying and labeling AI-generated content, ensuring that certain works, such as journalism, are marked as authentic or AI-generated.
Addressing AI Bias
To combat perceived bias against conservative figures in AI systems, Blackburn proposes that developers of high-risk AI conduct annual independent audits to detect political discrimination. It also outlines that federal agencies can only procure models adhering to unbiased AI principles.
AI Safety Measures
The TRUMP AMERICA AI Act introduces a baseline “duty of care” for AI chatbot developers, requiring them to take reasonable steps to mitigate foreseeable user harm. A risk-based regulatory framework for AI systems will also mandate developers to participate in evaluation programs.
Developers and deployers of AI will be held legally accountable for harm caused by their systems. Companies and federal agencies are required to report quarterly on AI-related job impacts, including layoffs or job displacements.
To address rising energy costs linked to AI data centers, Blackburn proposed safeguards ensuring that ratepayers are not unduly burdened by AI infrastructure expenses. This initiative aligns with Trump’s recent pledge to protect ratepayers from the costs associated with powering AI and data centers.
