AI technology is like a double-edged sword, offering convenience but also posing risks. To better manage these risks, Governor Gavin Newsom of California recently signed nine bills specifically targeting AI-generated content, especially the phenomenon of deepfakes.
Image source: Picture generated by AI, authorized service provider Midjourney
Among these bills, SB1047 is particularly noteworthy, requiring AI developers to implement safety measures to mitigate catastrophic events that the technology might trigger, such as mass casualties or cyber-attacks. Governor Newsom has not yet explicitly stated his stance on this bill. AB2602 and AB1836 focus on protecting the rights of actors and performers.
AB2602 stipulates that when using AI-generated voices or images of performers, a contract must be signed, and the performer must have a professional representative involved in the negotiations.
AB1836 prohibits the use of deepfake images of deceased performers in commercial film and television productions without the consent of their family members.
Newsom stated: "We have been exploring how AI and digital media change the entertainment industry, but our core focus has always been on protecting workers. These bills ensure that the industry can continue to thrive while strengthening the protection of workers."
The bills also include measures to combat the abuse of AI-generated content:
SB926 criminalizes the creation and distribution of seemingly real pornographic images;
SB981 requires social media platforms to establish reporting mechanisms for handling user-reported pornographic deepfake content;
SB942 mandates that AI systems include invisible watermarks in generated content and provide detection tools.
In election activities, AB2655, AB2839, AB2355, and AB2905 aim to combat the use of deepfakes and other deceptive digital content to protect the fairness and transparency of elections.
AB2655 requires major online platforms to remove or flag deceptive or digitally altered political content during specific periods before and after elections and to provide reporting methods.
AB2839 extends the time frame during which political entities are prohibited from distributing advertisements or other election materials containing deceptive AI-generated content.
AB2355 requires political advertisements using AI-generated content to disclose that the material has been altered.
AB2905 requires automated dial-up calls to inform recipients whether the voice is artificially generated.
Newsom emphasized: "Protecting the integrity of elections is crucial for democracy, and we must ensure that AI is not used to undermine public trust through misinformation. These measures will help combat the harmful use of deepfakes in political advertising."
Key points:
📜 Newsom signs nine bills to strengthen regulations on AI-generated content.
🎭 Protects the digital images of actors and performers, preventing unauthorized AI use.
🔍 Combats AI-generated pornographic content and deceptive information in election activities.