A Hawai‘i senator has introduced a federal bill that would require AI-generated content to be clearly labeled as such.
Sen. Brian Schatz co-introduced last week the AI Labeling Act as an effort to improve transparency as AI-generated content becomes more and more common in everyday life.
The measure that requires any AI-generated digital content — whether an image, video, audio or chatbot — include clear disclosures embedded within the content itself identifying it as AI-generated.
Those disclosures must also include a machine-readable disclosure, an invisible “stamp” embedded within the file’s metadata, that also describes the system that generated the content, and when it was generated.
All providers of generative AI systems would be required to ensure that their outputs’ disclosures are detectable by widely available AI-detection tools. Social media platforms would also be required to ensure that the disclosures are still visible when AI-generated content is viewed on their sites, and would be prohibited from obscuring or removing them.
The measure also includes a host of prohibitions against distributing AI content that doesn’t include the disclosures, or against adding disclosures to non-AI-generated content. People or entities who violate those prohibitions could be sued for up to $2,500 per violation.
“People deserve to know whether or not the videos, photos, and content they see and read online is real or original,” said Senator Schatz in a statement Tuesday. “Our bill is simple – if any digital content is made by artificial intelligence, it should be labeled so that people are aware and aren’t fooled or scammed.”
When introducing the bill, Schatz’ office specifically referenced a 2023 AI-generated photo of an explosion near the Pentagon, which triggered a dip in the stock market.
This wasn’t the only AI bill introduced last week. The AI Incident Reporting Act would require AI developers to report to the Secretary of Commerce any potentially harmful action autonomously taken by an AI model.
For instance, the bill requires that developers notify authorities if there is any sign that an AI model “is attempting to evade human oversight, deceive evaluators or operators … resist shutdown or modification … enable or accelerate the development, acquisition, or use of chemical, biological, radiological, nuclear or explosive weapons,” or other behaviors.
Neither bill has been voted on in Congress so far.
Meanwhile, the House voted last week to pass the Small Business Administration Artificial Intelligence Utilization Act, a measure requiring the SBA to publish regular reports about how it uses AI. No representative voted against the measure’s passage.
Other bills the House passed last week include:
• The Native American Entrepreneurial Opportunity Act, which establishes within the SBA an Office of Native American Affairs that would work to promote SBA programs with Native American tribes and Native Hawaiian organizations. The bill passed the House with no objections.
• The COVID Fraud Transparency Act, which requires the SBA to report to Congress about the number of COVID loan fraud cases that have been reported and resolved. The bill passed the House with no objections.
• The Emergency Conservation Program Improvement Act would allow agricultural producers to receive reimbursements for up to 75% of the cost of replacing farm or conservation structures following a natural disaster through the Emergency Conservation Program and Emergency Forest Restoration Program. The bill passed 368-19, with Hawai‘i Reps. Jill Tokuda and Ed Case in support. The bill has been enrolled to President Donald Trump’s desk.
For the latest news of Hawai‘i, sign up here for our free Daily Edition newsletter.
