AI Labeling Act of 2026

The AI Labeling Act of 2026 requires providers of generative AI systems and covered online platforms to label AI-generated digital content with clear disclosures and machine-readable provenance information, prohibits fraudulent tampering with those disclosures, and establishes enforcement mechanisms and a working group to set technical standards. Sec. 2, Required disclosures for covered AI-generated content: Providers of generative AI systems that produce covered AI-generated content must label it with a clear disclosure identifying it as AI-generated, embed a machine-readable disclosure containing details like the system version and creation date, ensure detection is possible without undue burden, and collaborate with covered online platforms. An exemption exists for content used solely for internal research and development not intended for public release. Covered online platforms must display such content with a clear label, not tamper with disclosures, and provide users options to share provenance information. Sec. 3, Protection of disclosures: It prohibits knowingly adding false disclosures to deceive, distributing AI content without required disclosures or adding such disclosures to non-AI content for financial benefit, and manufacturing or offering products primarily designed to circumvent or falsify disclosures. Exemptions allow individuals to read disclosures and apply to nonprofit libraries, archives, and educational institutions unless they act for commercial advantage or financial gain. Sec. 4, Enforcement by the Attorney General of the United States: The Attorney General may bring civil actions for violations of Sec. 3(a), with courts able to grant injunctions, impound devices, award statutory damages up to $2,500 per act or $25,000 per violation, triple damages for repeated violations within three years, and reduce damages for innocent violations. Sec. 5, Enforcement by States: State attorneys general may bring civil actions on behalf of residents for Sec. 3(a) violations, with relief similar to Sec. 4(c), but must notify the U.S. Attorney General and FTC, who may intervene. No duplicative compensatory damages are allowed if already awarded under Sec. 4 or 6. Sec. 6, Enforcement by private parties: Providers of generative AI systems or covered online platforms harmed by a Sec. 3(a) violation may sue for relief under Sec. 4(c), with notification to the Attorney General and FTC required, and no duplicative compensatory damages if already awarded under Sec. 4 or 5. Sec. 7, AI-generated content consumer transparency working group: Within 90 days, the Director of NIST, with the FTC, must establish a working group including federal agencies, AI developers, standards groups, platforms, academics, privacy advocates, media, technical experts, and victim representatives. The group will provide technical standards for labeling, ensure labels are tamper-resistant and interoperable, and advise the FTC on feasibility and detection criteria. Sec. 8, Definitions: Key terms are defined, including artificial intelligence chatbot, Commission (FTC), content provenance, covered AI-generated content (digital content materially altered by AI so a reasonable person would not believe it is AI-generated), covered online platform (with at least 10 million monthly users or $1.5 billion revenue), digital content, generative artificial intelligence system, liar's dividend, machine-readable, non-AI-generated content, and open-source.