Legislators in the U.S. filed a bill requiring creators of foundational AI models to disclose their sources of training data.
| Photo Credit: Reuters
Legislators in the U.S. filed a bill requiring creators of foundational AI models to disclose their sources of training data enabling copyright holders to know if their data was used.
The AI Foundational Model Transparency Act would direct the Federal Trade Commission to work in tandem with the National Institute of Standards and Technology (NIST) to establish rules for reporting training data transparency, a report from The Verge said.
Under the law, companies that make foundational models will be required to report the source of data used by them, along with information about efforts to prevent the models from providing inaccurate or harmful information.
The bill complements the Biden administration’s AI executive order, which provides rules around generative AI. These rules include the creation of standards for AI safety and security along with protecting privacy, advancing security and civil rights in the country.
(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)
Several lawsuits have been filed against companies developing and training AI models alleging copyright infringement.
In January, stock image and media content provider Getty Images announced it was suing Stability AI, the developers of the text-to-image deep learning model Stable Diffusion.
At the time the company claimed that Stability AI unlawfully copied and processed millions of images protected by copyright and associated metadata owned or represented by Getty Images.