On June 25, Senators Brian Schatz, John Curtis and Mark Warner reintroduced the AI Labeling Act. The bipartisan bill would require AI-generated audio, video and images to carry a disclosure. A 2023 version did not become law. That version was sponsored by Schatz and Republican John Kennedy.
Under the bill, generative AI providers would attach both a visible label and a machine-readable record of the system used and when the content was made. Large platforms, those with 10 million monthly US users or $1.5 billion in revenue, would have to show the labels and could not strip them out. Chatbots would have to say they are AI. The FTC would enforce the rules, and NIST would lead a group setting technical standards and looking at text. Schatz summed up the bill simply: content made by AI should say so, so people are not fooled or scammed. Warner said the US should catch up with other countries and set a global standard for disclosure and anti-circumvention.
Supporters include SAG-AFTRA, the Songwriters Guild of America, Music Creators North America, the Society of Composers and Lyricists, the Authors Guild, WGA East and IATSE. The National Association of Voice Actors said its members' voices are already being cloned without disclosure. Consumer groups such as Public Citizen also back the bill.
The bill does not mention streaming services by name. It lands alongside the NO FAKES Act, reintroduced in May with label and Spotify support, and as Tidal, Deezer, Apple Music and Spotify each roll out their own ways of flagging AI tracks. Tidal said the day before MBW's report that it would tag fully AI tracks and stop paying royalties on them. Spotify has also said profiles that mainly represent AI or AI-persona acts cannot get its new verification badge.
Sources · 1