As part of Congressman Lieu's Youth Advisory Council, I have gained the opportunity to write a bill proposal about a topic that I care about. After listening to the Daily by the New York Times episode titled "The Writer's Revolt Against A.I. Companies", I decided to write my bill about the negative impact AI technologies have on creators. The purpose of my bill, the Artificial Intelligence Content Disclosure Act, is to institute guidelines for disclosing the use of artificial intelligence generated content in creative works, informing consumers of possible bias and inaccurate information.
After finishing my proposal, the Council voted on all the bill proposals and decided that mine was one of the 3 that should be further developed. I worked with Congressman Lieu's legal team to make sure that it was feasible and constitutional, and on adjusting the wording of the bill. At the end of the year, my bill was the only one that could reasonably be introduced in the House of Representatives, and the Congressman worked to get my bill introduced.
Here is the contents of my proposal:
To: Congressman Ted Lieu
From: Youth Advisory Council
RE: Artificial Intelligence Content Disclosure Act
Bill Summary and Specifics
The purpose of the Artificial Intelligence Content Disclosure Act is to institute guidelines for disclosing the use of artificial intelligence (AI) generated content in creative works, informing consumers of possible bias and inaccurate information. Creative works can be defined as anything requiring creative effort, which includes but is not limited to films, articles, stories, music, and artwork. This bill will require all publishers of creative works to clearly disclose the use of artificial intelligence in their work. Specifically, these requirements would include:
Prominently displayed disclosure of AI usage
Header or footer of written works
Overlay on digital content
Credits in films, videos, and other visual works
Labels on physical artworks
Information about the specific AI model used
Information, if possible, about any possible biases the model contains