Artificial intelligence in art: innovation or dispossession?

The voice of legendary voice actor Pepe Lavat, who died in 2018, was artificially resurrected for an institutional message from the National Electoral Institute. The scandal symbolizes how the emergence of artificial intelligence (AI) in the performing arts and entertainment has sparked a confrontation between digital progress and labor dignity.

As synthetic voices reach superlative levels of realism, voice actors, screen performers, and broadcasters face the possibility of their voices and images being replicated without their consent and their artwork becoming raw material for algorithms.

The transformative capabilities of AI are evident in the audiovisual sector. The technology enables voice cloning with astonishing accuracy. It captures the timbre, cadence, and inflection that make each voice unique. It rejuvenates voices, resurrects classic performances, or modifies existing ones without the need for re-recordings. Notable examples include recreating the voice of a young Luke Skywalker in The Mandalorian and authenticating Hungarian dialogue in The Brutalist .

Instead of a one-time performance, an actor's voice or image becomes a persistent data asset that can be manipulated and reused endlessly. The question is whether the AI-generated output is a new "performance" requiring new compensation, or simply a "data use" derived from a past performance. The distinction is crucial for defining compensation models and consent mechanisms.

The digital resurrection of deceased actors and the potential for AI to perform a role unrelated to their real-life counterparts introduces the concept of "perpetual performance." An artist's image or voice can continue to "perform" after their physical career ends or even after their death. Who owns the rights to a posthumous digital image?

Rapid advances in Generative AI have been driven by vast datasets, often acquired without the consent of or compensation to the original creators. The practice has led to accusations of intellectual property theft that have shaken copyright laws.

Republican Senator Josh Hawley has declared that training AI models with copyrighted works constitutes "the largest theft of intellectual property in American history." It represents "criminal conduct" rather than aggressive business tactics. Hawley dismisses the argument that such practices are necessary to win the AI race against China.

Major AI companies have faced lawsuits for using copyrighted data and content to train their models. Major media conglomerates have sued AI companies for copyright and trademark infringement.

The direct impact of AI on artists has mobilized unions and guilds seeking strong protections, but questions arise about the ultimate beneficiaries of such regulation. Voice actors oppose having their voices cloned and sold without their consent. They demand explicit consent and fair compensation for any use of their digital voice replicas.

The Mexican Association of Commercial Announcers (AMELOC) protested against AI voice cloning and demanded legal recognition of voices as biometric identifiers. The demand arose after the INE (National Institute of Statistics) unauthorized use of the late actor José Lavat's voice. The National Actors Association (ANDA) has proposed legislative initiatives to protect voice actors from AI. They warn that, without regulation, human dubbing could disappear, leading to job losses.

There is a risk that associations like ANDA, AMELOC, and AITE will be the primary beneficiaries of regulation, rather than individual artists. If regulation focuses on collective bargaining or licensing agreements managed by these guilds, it could centralize control and revenue streams for them, not for the creators.

As the impact of AI becomes undeniable, countries are beginning to develop regulatory responses. California has enacted laws designed to protect artists. They prohibit the creation of AI-generated replicas of artists without consent and grant the heirs of deceased actors the right to sue for unapproved uses of their likeness.

Denmark is seeking to combat deepfakes, which are realistic digital imitations of an artist's performance or personal characteristics (face, body, voice) without consent. The goal is to allow individuals to request the removal of such content from digital platforms, with the possibility of those platforms being subject to fines for non-compliance.

The European Union's AI Law includes provisions for developers of Large Language Models (LLMs). It strengthens copyright enforcement, requires general-purpose AI providers to implement policies that comply with EU copyright law, and includes respect for opt-out mechanisms for text and data mining.

The trend in legislation is toward proactive measures rather than relying on post-infringement litigation. California's consent and Denmark's emphasis on deletion requests give individuals initial control. EU transparency for training data and opt-out mechanisms seek to prevent unauthorized use before it occurs.

Mexico's more than 60 legislative attempts since 2020 have not translated into concrete legal frameworks. The absence of a national AI strategy is a key impediment. Bills pursue the ethical and safe development of AI through privacy, cybersecurity, and neuro-rights. Some bills propose a risk-based approach copied from the EU, classifying AI systems as "Unacceptable Risk," "High Risk," and "Low Risk." There are also proposals to amend the Constitution to grant Congress explicit authority to legislate on AI.

The path forward for AI regulation in the arts must balance safeguarding human creativity, preserving jobs, and fostering innovation. Regulating AI out of ignorance, fear, or in anticipation of technological advances carries more risks than benefits.

Overregulation could hinder technological innovation, restrict the emergence and development of new businesses, limit economic growth, and reduce competition. It could also lead to a homogenization of creative expression, eroding the unique human touch, intuition, and emotional depth that define art.

___________________________________________

Twitter: @beltmondi