Gavin Newsom Vetoes AI Safety Bill, Which Had Backing of SAG-AFTRA

As a lifelong movie buff with a keen interest in the intersection of technology and entertainment, I find myself deeply troubled by Governor Newsom’s veto of SB 1047. The potential threats posed by advanced AI models are not just theoretical; they are real, and we’ve seen glimpses of them in our industry already.

Governor Gavin Newsom of California rejected a proposal on Sunday, aimed at mitigating potential disasters caused by sophisticated artificial intelligence systems.

During the legislative session, SB 1047, a highly debated AI legislation, garnered significant attention. On both sides stood numerous activists and AI corporations. The creators of the bill expressed concern that without regulation, AI models might be employed for creating chemical or nuclear weapons, posing a threat to countless lives. However, opponents contended that excessive regulations could hinder the progress of AI, potentially causing AI companies to depart from the state.

In his veto message, Newsom expressed that the bill tackles a legitimate issue, yet it lacks an appropriate regulatory system.

He stated that this method isn’t likely the most effective way to safeguard the public against genuine risks associated with technology. In essence, any system for governing AI should adapt as swiftly as the technology does.

Supporting this legislation was SAG-AFTRA, the union that advocates for Hollywood actors. On Tuesday, “Artists for Safe AI” – a group backing artificial intelligence safety – published an open letter endorsing the bill as well. Notable figures such as J.J. Abrams, Shonda Rhimes, Judd Apatow, Rob Reiner, Jane Fonda, Rian Johnson, Adam McKay, Mark Hamill, Mark Ruffalo, Don Cheadle, and others have lent their backing to this cause.

The acting guild has voiced concerns about the possibility of AI duplicating actors for films and television productions without their permission. This is a new stance they’ve taken, addressing broader potential hazards beyond the entertainment industry that might arise from sophisticated artificial intelligence technologies.

Jeffrey Bennett, the union’s general counsel, stated that the basis for their support lies in our direct encounter with the risks associated with one facet of AI. The proposed bill appears to be unique in addressing only the highly advanced, costly systems that possess the potential to trigger a widespread critical issue. Isn’t it more prudent to control at this level? Why not incorporate common-sense safety mechanisms at this stage itself?

As a passionate movie-goer, I eagerly awaited the governor’s announcement on Sunday, where he expressed his intention to veto certain measures. Yet, his message didn’t end there; he also promised to gather leading experts to establish guidelines that ensure the safe evolution of Artificial Intelligence. Rest assured, this isn’t a one-year endeavor for him – he plans to keep working on it in the coming year as well.

This year, SAG-AFTRA backed two legislative measures regarding AI usage within the entertainment sector in California. Governor Newsom recently approved these bills during a ceremony held at the SAG-AFTRA headquarters.

The union chose not to participate in the discussion about SB 1047 when it was under consideration in the Legislature. However, on September 9th, they wrote a letter to Newsom advocating for his approval of the bill.

As a passionate advocate, I’ve observed how AI-powered deepfake technology has been manipulated to fabricate explicit and nude images featuring SAG-AFTRA members. In my opinion, it’s crucial for policymakers to take action and safeguard both our union members and the public. SB 1047 is a thoughtful initial move that paves the way towards this protection.

Other Hollywood unions and businesses chose not to intervene in this matter, as it specifically concerned advanced AI models that are yet to be developed.

Governor Newsom has approved another legislation, AB 2013, that mandates AI model developers to openly declare if they utilize copyrighted materials during their training process.

The Concept Art Association actively backed this legislation. This organization is a representative body for artists who generate the visual concepts for films, animations, and video games. Lately, their creations have often been incorporated into artificial intelligence models, which may potentially replace their roles in the industry.

From a movie-loving perspective, I’d rephrase it like this: When it comes to developing models, there’s no rule that forces us to disclose the complete dataset we’ve used for training. Also, there’s no obligation to pay for using copyrighted works, an issue that’s still under debate in courts. However, what the rule does insist on is that we should credit or acknowledge the use of copyrighted data, or any personal information, whenever it’s applicable in our work.

Deana Igelsrud, from the Concept Art Association, stated that any information revealed is beneficial,” she said. “Though it’s quite broad, it’s a step in the right direction.” (This version maintains the original meaning and flows smoothly.)

Representative Adam Schiff’s proposed bill in Congress has received support from the Hollywood unions, an initiative he put forth earlier this year.

Igelsrud stated, “These AI systems wouldn’t produce anything without being loaded with art from throughout history. I believe many people underestimate the fact that real human beings are connected to this data. People often assume that if something is on the internet, it can be used freely. However, this is not the case.

2024-09-30 03:48

Gavin Newsom Vetoes AI Safety Bill, Which Had Backing of SAG-AFTRA

Read More

Minority Mindset