This bill, the "SIMS Act," establishes a federal offense for providing chatbots that simulate minors engaging in sexually explicit conduct or conversation.
Bill Cassidy
Senator
LA
This bill, the "Stopping Illegal Minor Simulations Act" (SIMS Act), prohibits the creation and provision of chatbots that simulate minors engaging in sexually explicit conduct or conversation. It establishes federal criminal and civil penalties for covered entities that willfully violate this ban. The law includes exemptions for research and law enforcement investigations.
The 'Stopping Illegal Minor Simulations Act' (SIMS Act) creates a new federal offense to shut down the development and operation of chatbots designed to mimic minors in sexually explicit scenarios. Under this bill, any person or company providing a chatbot that simulates a minor engaged in interactive, sexually explicit conduct or conversation—provided the content is legally obscene—faces significant legal heat. This isn't just about static images; it’s specifically targeting the new wave of AI that can hold a conversation, respond to natural language, and mimic human emotions. The bill sets a clear boundary: if you own or operate a 'covered entity' that makes these specific types of AI available to users in the U.S., you are on the hook.
The bill defines a 'chatbot' broadly enough to cover everything from website-based AI to standalone software apps that produce 'adaptive or context-responsive output.' For a tech developer or a small software firm, this means the legal 'fine print' now includes a strict $100,000 fine for willful violations. The legislation specifically ties its definition of 'sexually explicit conduct' to existing federal child exploitation laws (18 U.S.C. 2256), ensuring that the standards for what constitutes a crime are consistent with current laws protecting real children. Whether the simulated minor is based on a real person or is entirely AI-generated doesn't matter—the simulation itself is the violation.
To make sure this doesn't accidentally tank legitimate AI safety work, the bill includes a 'red-teaming' carve-out. This means if you are a cybersecurity researcher or a developer testing a system to see if it can be manipulated into breaking these rules, you aren't liable for that internal evaluation. Additionally, the Attorney General can grant specific exemptions for law enforcement agencies—local, federal, or even certain international partners—who might need to use these tools as part of undercover investigations into actual child exploitation. It’s a move designed to keep the tools out of the hands of the public while letting the 'good guys' use the tech to catch real-world predators.
Once this bill passes, the clock starts ticking: the rules and penalties go live exactly 180 days later. Beyond the $100,000 criminal fine, the Attorney General is empowered to hit violators with civil lawsuits to freeze operations and demand restitution. To keep the process transparent, the Department of Justice will have to hand over an annual report to Congress. This report will detail exactly how many people and corporations were convicted, how much money was collected in fines, and how often law enforcement used these chatbots in their own investigations. For the average person, this means a significant shift in how the government monitors the 'darker' corners of AI development, focusing heavily on preventing the normalization of simulated exploitation.