Blab AI
Senator Schatz Proposes Federal AI Labeling Bill to Mandate Disclosure of AI-Generated Content
A Hawaii senator has introduced a federal bill that would require any content created by artificial intelligence to carry a clear label. Senator Brian Schatz co‑introduced the AI Labeling Act last week, arguing that the growing use of generative AI tools makes it harder for people to know whether a photo, video, audio clip or chatbot response is real or machine‑generated.
The bill would apply to all digital media produced by AI systems, including images, videos, audio recordings and text generated by chatbots. It requires two types of disclosure. First, a visible label must appear in the content itself, such as a banner or caption that says the material was created by AI. Second, a machine‑readable “stamp” would be embedded in the file’s metadata. The stamp would identify the system that produced the content and the date of generation.
Under the proposal, AI developers would have to ensure that their outputs can be detected by widely available AI‑detection tools. Social media platforms would also be required to keep the visible label on any AI‑generated content that appears on their sites, and they would be prohibited from removing or hiding the label.
The bill includes penalties for non‑compliance. Distributing AI content without the required disclosures, or adding a label to content that was not AI‑generated, could result in a civil lawsuit. The lawsuit could seek up to $2,500 for each violation.
Schatz cited a 2023 incident in which an AI‑generated photo of an explosion near the Pentagon was circulated online and caused a temporary dip in the stock market. The example was used to illustrate how misinformation can spread quickly when people cannot distinguish real from synthetic media.
The AI Labeling Act is one of several AI‑related bills introduced in Congress last week. The House passed the Small Business Administration Artificial Intelligence Utilization Act, which requires the SBA to publish regular reports on its use of AI. The House also approved the Native American Entrepreneurial Opportunity Act, the COVID Fraud Transparency Act, and the Emergency Conservation Program Improvement Act.
In the Senate, another bill was introduced that would require developers of high‑risk AI models to report incidents to the Secretary of Commerce. The AI Incident Reporting Act would trigger a notification if an AI system attempts to evade human oversight, deceives operators, resists shutdown, or could facilitate the creation of weapons.
None of the bills have been voted on yet. The AI Labeling Act is still in the early stages of the legislative process, and it will need to pass through committee hearings and floor debates before it can become law.
The proposal reflects a broader trend of governments seeking to regulate the transparency and safety of generative AI. Similar measures have been discussed in the European Union and in several U.S. states, where lawmakers are exploring ways to make AI‑generated content identifiable and to prevent the spread of misinformation.
At this point, the AI Labeling Act remains a proposal. Its future will depend on congressional support, industry feedback, and the evolving public debate over how best to balance innovation with accountability. If enacted, the bill would become the first federal requirement that AI‑generated media be clearly labeled, potentially setting a precedent for other jurisdictions.
The bill would apply to all digital media produced by AI systems, including images, videos, audio recordings and text generated by chatbots. It requires two types of disclosure. First, a visible label must appear in the content itself, such as a banner or caption that says the material was created by AI. Second, a machine‑readable “stamp” would be embedded in the file’s metadata. The stamp would identify the system that produced the content and the date of generation.
Under the proposal, AI developers would have to ensure that their outputs can be detected by widely available AI‑detection tools. Social media platforms would also be required to keep the visible label on any AI‑generated content that appears on their sites, and they would be prohibited from removing or hiding the label.
The bill includes penalties for non‑compliance. Distributing AI content without the required disclosures, or adding a label to content that was not AI‑generated, could result in a civil lawsuit. The lawsuit could seek up to $2,500 for each violation.
Schatz cited a 2023 incident in which an AI‑generated photo of an explosion near the Pentagon was circulated online and caused a temporary dip in the stock market. The example was used to illustrate how misinformation can spread quickly when people cannot distinguish real from synthetic media.
The AI Labeling Act is one of several AI‑related bills introduced in Congress last week. The House passed the Small Business Administration Artificial Intelligence Utilization Act, which requires the SBA to publish regular reports on its use of AI. The House also approved the Native American Entrepreneurial Opportunity Act, the COVID Fraud Transparency Act, and the Emergency Conservation Program Improvement Act.
In the Senate, another bill was introduced that would require developers of high‑risk AI models to report incidents to the Secretary of Commerce. The AI Incident Reporting Act would trigger a notification if an AI system attempts to evade human oversight, deceives operators, resists shutdown, or could facilitate the creation of weapons.
None of the bills have been voted on yet. The AI Labeling Act is still in the early stages of the legislative process, and it will need to pass through committee hearings and floor debates before it can become law.
The proposal reflects a broader trend of governments seeking to regulate the transparency and safety of generative AI. Similar measures have been discussed in the European Union and in several U.S. states, where lawmakers are exploring ways to make AI‑generated content identifiable and to prevent the spread of misinformation.
At this point, the AI Labeling Act remains a proposal. Its future will depend on congressional support, industry feedback, and the evolving public debate over how best to balance innovation with accountability. If enacted, the bill would become the first federal requirement that AI‑generated media be clearly labeled, potentially setting a precedent for other jurisdictions.
