A federal proposal aims to label AI-generated content, impose liability for harmful deepfakes, and allow employees to report AI risks without retaliation.

The American Leadership in AI Act, introduced by Rep. Ted Lieu (D-Calif.) and Rep. Jay Obernolte (R-Calif.), seeks to establish federal rules on harmful AI-generated deepfakes and new protections for employees who report AI-related risks.

The proposed bill has four core components: 

1. Curtail harmful deepfakes. Under the proposal, it would be unlawful to knowingly create or distribute harmful deepfakes, including content that falsely represents a person in a way that could cause reputational, financial, or personal harm. Further, online platforms that knowingly host such content or fail to act on it after being notified would be liable.
2. Disclose AI-generated content. AI-generated content that could be mistaken for real content would need to include clear and conspicuous labeling to help users distinguish authentic content from manipulated material.
3. Protect whistleblowers. The proposal would prohibit AI companies from retaliating against employees who report concerns related to AI systems, including safety risks, legal violations, or failures to comply with federal standards. Whistleblowers would be able to take legal action if they face retaliation, including reinstatement, compensation, and other damages.
   1. Whistleblower reporting mechanisms. Employees would report issues internally, to federal regulators, or even to Congress. In addition, companies would be restricted from using nondisclosure agreements to prevent employees from reporting AI-related risks to authorities. The bill directs federal agencies to establish processes for receiving and reviewing whistleblower disclosures related to AI systems.
4. Greater legislative coordination. These include support for international standards-setting efforts and coordination across federal agencies on AI policy and development.

Backdrop

To date, states have been outpacing the federal government in enacting AI legislation. Some, like California’s SB-53, are more far-reaching. The Trump administration seeks to limit state laws in favor of a more ubiquitous federal framework. The White House introduced its seven-point recommendation back in March. The latest proposal from members of Congress is a renewed effort to establish federal AI guidance.