California Governor Gavin Newsom signs Deepfake and other AI bills into law (22.09.2024)

Published on: 22nd September, 2024 | Authored by: Mr Archak Das

Key Highlights:

Combatting Deepfake Nudes, Election Manipulation and other issues: Governor Newsom recently, over the last week, has signed critical laws to criminalize AI-generated deepfake nudes, tackle AI’s role in electoral manipulation and other AI-related issues.
California Leads in AI Regulation: California, home to many of the world’s top AI companies, is at the forefront of addressing AI-related issues with groundbreaking laws. Governor Gavin Newsom has already signed eight AI-related bills into law, with more in the pipeline, making the state a leader in AI governance.
A Pivotal AI Bill: Senate Bill 1047, currently awaiting Governor Newsom’s decision, is a landmark piece of legislation requiring AI developers to implement safeguards against catastrophic harm from advanced AI models. While controversial, it could shape the future of AI regulation.

California, as the global epicenter of AI innovation, is making a bold move by implementing laws that address the societal impacts of AI. Governor Gavin Newsom is currently considering 38 AI-related bills, including the highly debated SB 1047, aimed at safeguarding against the dangers of advanced AI systems.

The Deepfake Dilemma

One of the most alarming applications of AI has been the creation of deepfake nudes, where AI-generated images of individuals without their consent were circulated and such technology can be weaponized for blackmail, harassment, and reputation damage. Recognizing this threat, Governor Newsom signed two crucial bills into law on 19^th September, 2024 and these reflect California’s strong stance on privacy protection, as they hold perpetrators accountable and force platforms to be more vigilant:

SB 926: Criminalizes the act of blackmailing individuals using AI-generated nude images that resemble them and this landmark law aims to deter individuals from exploiting AI in this harmful way.
SB 981: Mandates that social media platforms create mechanisms for reporting deepfake nudes, where platforms must investigate reported content and take down any confirmed AI-generated nudes, ensuring user protection from malicious AI usage.

AI-Generated Content Disclosure

A significant concern with AI technologies is the seamless blending of AI-generated and real content and to address this, California has introduced SB 942, signed by Newsom on 19^th September, 2024 which mandates that AI systems disclose when content is AI-generated. This law requires AI-generated content to include watermarks or metadata that clearly indicate its AI origins. For example, images produced by AI tools like DALL-E must now contain information in their metadata revealing that they were AI-generated as the goal of this law is transparency. As AI tools become increasingly powerful, they blur the line between reality and fiction and by mandating that AI-generated content is labeled, California hopes to help users identify such content and prevent its misuse in areas like journalism, entertainment, and social media.

Election Integrity

Another crucial area of concern is the potential for AI to disrupt democratic processes. AI-generated deepfakes can be used to spread misinformation during elections, swaying public opinion and misleading voters. So, three new laws signed by Governor Newsom directly address this issue as these aim to preserve the integrity of the electoral process in the age of AI, recognizing the severe consequences of unchecked AI influence on democracy:

AB 2655: Requires large social media platforms to remove or label AI-generated deepfakes related to elections and failure to comply allows candidates to seek injunctive relief.
AB 2839: Targets individuals who post or repost AI deepfakes intended to deceive voters and such law was enacted immediately, and even high-profile individuals, like Elon Musk, may be affected by this.
AB 2355: AI-generated political advertisements must now disclose that the content is AI-generated and this helps ensure transparency in political advertising, making it harder for voters to be misled by AI-generated content.

Actors vs AI

The entertainment industry has also been affected by AI, as studios explore creating digital replicas of actors, sometimes without their consent and with concerns growing over the use of AI to generate lifelike representations of actors, California has stepped in to ensure performers’ rights are protected. Governor Newsom signed two laws on 17^th September, 2024 that focus on this issue and these have been championed by SAG-AFTRA, the nation’s largest actors’ union, as they help establish clear boundaries around the use of AI in the entertainment industry:

AB 2602: Requires studios to obtain explicit permission from actors before creating AI-generated replicas of their likeness or voice.
AB 1836: Prohibits studios from creating digital replicas of deceased actors without the consent of their estates, safeguarding the legacy of performers.

The Senate Bill 1047

The most contentious AI-related bill on Governor Newsom’s desk is Senate Bill 1047, also known as the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act. This bill is at the center of a broader debate on how best to regulate AI while promoting innovation as it requires AI developers to integrate safeguards into advanced AI models to prevent catastrophic harms, such as large-scale privacy violations, or even existential threats posed by runaway AI systems. The California attorney general would have the power to enforce these regulations, holding developers accountable if they fail to take reasonable steps to prevent harm.

However, the bill has met with significant opposition from AI companies, who argue that:

The bill’s definition of “covered models” is too rigid and doesn’t allow for technological advancements.
It’s unfair to hold AI developers responsible for how others might misuse their technology.
The bill could stifle innovation, particularly for smaller AI startups that may struggle to meet compliance requirements.

Despite these concerns, proponents argue that regulating AI is essential to prevent potential disasters and a veto on SB 1047 could signal reluctance to impose any meaningful restrictions on AI technologies, leaving society vulnerable to unchecked AI risks. Governor Newsom’s decision on SB 1047, due by the end of September, will have significant implications for the future of AI regulation in the United States.

Conclusion

California’s proactive approach to AI regulation is shaping the future of how this powerful technology will be governed and from combatting deepfake nudes to ensuring election integrity, the state is taking bold steps to address the immediate risks posed by AI. SB 1047 stands out as a pivotal moment in this effort, as its success or failure could determine how far the state and the nation is willing to go in regulating AI. As AI continues to evolve, California’s actions will serve as a blueprint for other governments, offering valuable lessons on how to navigate the complex balance between innovation and safety.