H.R. 9914 · Collaboration on Adversarial Threats and Security Risks Act
If AI companies agree between themselves to slow down, they may be in violation of antitrust laws. The CATS Act would give them an exemption to share information and agree to take joint actions to avoid AI risks. For example, they would be allowed to slow down or pause development.
H.R. 9477
Right now, if a severe incident occurs with AI, nothing forces the companies to report what happened. We already have cases where companies did not disclose significant incidents. Without disclosure, external parties cannot track how serious issues with AI have become.
The AI Incident Reporting Act would require covered AI developers to report serious safety incidents and dangerous capabilities to the Department of Commerce. These reports would inform government oversight; however, the bill does not require them to be made public.
We support this act because government needs to know when AI systems show dangerous capabilities or escape safeguards so government can respond before the risks grow.
H.R. 3460 / S. 1792
If an employee reports a serious issue about their AI company, the company might attempt to punish them through actions like firing, suspending, or demoting the employee. People working inside AI companies need to be able to raise serious concerns without fear of punishment or retaliation.
Whistleblower protections make it illegal for companies to retaliate against employees who report AI-related legal violations, security vulnerabilities, or failures to address substantial dangers to public safety, public health, or national security through protected channels.
H.R. 3447 / S. 1705
To create powerful AI that is dangerous, AI companies need access to advanced AI chips. Currently there are restrictions on countries like China and Russia that limit the advanced AI chips they can buy. However, these restrictions can be bypassed through illegal smuggling.
The Chip Security Act would require technical safeguards to help detect and prevent unauthorized countries and users from smuggling chips or using other methods to gain the chips when they should not be able to.
S. 1473 / H.R. 6322
The Stop Stealing our Chips Act would create financial rewards and protections for whistleblowers who help expose export-control violations such as smuggling chips to countries like China and Russia that should not have access.
It has a similar goal to the Chip Security Act, but its purpose is to reward and protect people who report chip-smuggling networks.
H.R. 8283
Foreign actors, including those in China and Russia, can try to copy powerful AI made by US companies by extracting their capabilities without permission.
The Deterring American AI Model Theft Act would establish a confidential channel for companies to report these attacks, require the Commerce Department to consult companies and experts on defenses and publish security guidance, and authorize sanctions against foreign actors involved in these attacks.
These measures aim to make unauthorized copying harder and limit the spread of powerful AI capabilities to actors who would evade safeguards and export restrictions.
H.R. 10362
Federal agencies are increasingly using advanced AIs. The Stop Rogue AI Act aims to protect federal agencies from rogue AIs by directing the National Institute of Standards and Technology (NIST) to develop standards for identifying, monitoring, and controlling AI agents on computer networks, and requiring federal agencies and contractors to build these safeguards into how they buy and deploy AI agents.
As AI agents take actions on our behalf, federal agencies and contractors need ways to see what they are doing, limit their access, and keep control.