2 months ago
Sen. Marsha Blackburn (R-TN) has unveiled a draft legislative framework that targets immoderate of nan astir basking fastener issues swirling astir Hollywood and AI, including nan usage of copyrighted useful successful training models and nan prevalence of unauthorized deepfakes of awesome celebrities.
Blackburn’s proposals besides telephone for sunsetting Section 230, nan proviso of a 1996 rule that exempts platforms from liability for 3rd statement content. The tech manufacture has fiercely lobbied against efforts to modify aliases abolish Section 230, arsenic it is considered a cornerstone of Silicon Valley’s maturation complete nan past 30 years.
Blackburn’s proposals besides see safeguards for minors, arsenic good arsenic provisions designed to “prevent favoritism based connected governmental affiliation.”
Blackburn said successful a statement, “Instead of pushing AI amnesty, President Trump rightfully called connected Congress to walk national standards and protections to lick nan patchwork of authorities laws that has hindered AI innovation.”
Blackburn has named nan authorities The Republic Unifying Meritocratic Performance Advancing Machine intelligence by Eliminating Regulatory Interstate Chaos Across American Industry, aliases nan TRUMP AMERICA AI Act.
Although nan sanction is simply a motion toward nan president’s telephone for national AI standards, Trump has pushed backmost against astatine slightest 1 of nan provisions, arsenic his management has taken a much lenient attack to AI development.
One of nan provisions successful Blackburn’s connection would make clear that nan unauthorized “reproduction, copying aliases computational processing of copyrighted works” successful training models is not a adjacent usage nether nan Copyright Act. That rumor is beforehand and halfway successful a number of lawsuits being brought by contented creators, arsenic AI giants contend that their usage falls extracurricular nan scope of copyright liability.
Trump has seemingly sided pinch tech connected nan issue, telling nan AI Summit successful Washington past year, “You can’t beryllium expected to person a successful AI programme erstwhile each azygous article, book aliases immoderate you’ve studied you’re expected to salary for. We admit that, but you conscionable can’t do that because it’s not do-able.”
Blackburn’s connection besides incorporates nan No Fakes Act, which would springiness individuals nan correct to power nan usage of their integer likeness. The measure was introduced past twelvemonth on pinch Sen. Chris Coons (D-DE), arsenic good arsenic a agelong database of supporters successful nan manufacture including SAG-AFTRA, studios and grounds labels, arsenic good arsenic tech companies for illustration Google and OpenAI.
The draught authorities besides includes different measure Blackburn has championed, nan Kids Online Safety Act, which would require platforms to “exercise reasonable care” successful designing features that whitethorn summation children’s online activity. It besides places restrictions connected nan usage of minors’ information and requires devices truthful parents tin entree childrens’ privateness settings.
Other aspects of nan connection see requirements that nan National Institute of Standards and Technology create standards for “content provenance information, watermarking and synthetic contented detection.” Another request would require that AI companies let owners of imaginative aliases journalistic contented to connect root information.
Other provisions require national agencies to study nan occupation effects of AI, and for nan Department of Energy to cod information connected nan likelihood of adverse AI incidents. The authorities besides would found nan National Artificial Intelligence Research Resource, designed to make computing resources, datasets and infrastructure disposable to students, researchers, non-profits, mini businesses and world institutions.
