AI regulation in California: an innovative and protective approach.

California, the cradle of technological innovation, has just taken a decisive step in the legal framework of artificial intelligence (AI) with the adoption of two major laws in September 2024.

These legislative texts, Assembly Bill 2602 (AB2602) and the California AI Transparency Act (SB942), establish a pioneering legal framework that aims to reconcile technological innovation and the protection of individual rights.

This legislative advance demonstrates the Californian state's will to address the ethical and legal challenges posed by the rapid development of AI, particularly in the field of digital replicas and generative AI.

I. Assembly Bill 2602: Protection of Artists against Digital Replicas.

A. Context and objectives of the law.

Assembly Bill 2602, effective January 1, 2025, is part of an initiative to protect the rights of performers in the face of emerging digital replication technologies.

This law recognizes the specific challenges associated with the use of digital replicas in the entertainment industry and establishes a strict legal framework for their use.

B. Definition and scope.

The law defines a \" _digital replica_\" as a computer-generated electronic representation that is highly realistic and readily identifiable as the voice or visual appearance of an individual.

This definition encompasses sound recordings, images, audiovisual works, or transmissions in which the individual did not actually perform or appear, or in which the fundamental character of the performance has been materially altered.

C. Main provisions.

The law establishes strict conditions for the use of digital replicas.

The central element of this regulation is the requirement for a specific contractual description of the intended uses of the digital replica. This provision aims to ensure transparency and protect artists' rights over their image.

The law also stipulates that the use of these digital replicas is contractually possible only if the performer cannot himself perform the said performances.

Finally, a crucial aspect of the law consists of the obligation of legal representation.

The law requires that the individual be represented, in the context of signing a contract, by an attorney negotiating the digital replica rights on their behalf, or represented by a labor union.

II. The California AI Transparency Act: towards increased transparency for Generative AI.

A. Objectives and scope.

Senate Bill 942 (SB942), which will come into effect on January 1, 2026, establishes transparency obligations for generative AI providers.

This law aims to create a framework of trust around the use of AI by imposing measures for the detection and identification of artificially generated content.

B. Scope of application.

The law applies to \" _covered providers_\", defined as individuals creating, coding, or producing generative AI systems with more than one million monthly visitors or users and publicly accessible within the geographical limits of the State.

C. Main obligations.

Providers must make available, free of charge, a tool to assess whether content (image, video, or audio) has been created or modified by their generative AI system.

Providers must also allow for the identification of generated content (image, video, audio) through an explicit mention and through metadata allowing knowledge of the originating AI system.

Conclusion.

The adoption of these two laws marks a significant turning point in AI regulation.

California sets an important precedent by creating a legal framework that seeks to reconcile technological innovation and the protection of individual rights. These laws could serve as a model for other jurisdictions seeking to regulate AI in a balanced and effective manner.

This regulatory approach demonstrates a commitment to maintaining California's leadership position in technological development while ensuring adequate protection of individual rights and transparency. It also illustrates the ability of legislators to adapt to rapid technological developments and propose innovative legal solutions.

The effectiveness of these laws and their real impact on industry and society can only be fully evaluated after their respective entry into force in 2025 and 2026.

Nevertheless, they already constitute a significant advance in AI regulation and establish standards that could influence future legislation in this constantly evolving field.