AI Is Not Waiting: Why Lawmakers Must Act Now

editorialAI
Written By Nina Katashvili

A policy analysis on why unregulated AI threatens cultural survival and why the music industry is already the proof.

Artificial intelligence is already transforming society for better and, without guardrails, for far worse. The real question is not whether AI will disrupt culture, labor, or truth. It already has. The question is whether laws, rights, and public protections will shape that transformation, or whether we will be left cleaning up the wreckage of cultural exploitation and economic collapse after it happens.

The Center for AI Safety demonstrates what responsible AI development can look like and why urgent, human-centered regulation is essential. From technical safety research to public-interest infrastructure, their work underscores one critical reality: the risks we face are systemic, present, and accelerating.

AI IS NOT JUST TECHNOLOGY. IT IS INFRASTRUCTURE.

AI is infrastructure embedded into nearly every layer of modern life.

  • Cultural production and creative labor

  • Information systems and political discourse

  • Economic power and job markets

  • Privacy, identity, and human rights.

Unlike cars, pharmaceuticals, or financial systems, AI operates with few binding regulations governing its development, deployment, or accountability.

The Center for AI Safety focuses on technical safety, scalable oversight, and transparency mechanisms that make risk measurable and governance possible. Their work makes one thing clear: policy delay is not neutrality. It is a decision to let harm scale unchecked.

WHAT THE CENTER FOR AI SAFETY DOES AND WHY IT MATTERS

The Center for AI Safety develops transparency tools, reliability benchmarks, and governance frameworks that can inform enforceable policy. In 2024, the organization introduced circuit breakers and other technical safeguards aimed at reducing large-scale societal risks from advanced AI systems.

AI safety cannot remain a closed corporate discipline. CAIS invests in education, collaboration, and open research to ensure oversight capacity exists outside Big Tech. Its programs expanded significantly in 2024.

CAIS also operates a shared compute cluster dedicated to independent AI safety research, preventing critical infrastructure from being monopolized by corporations alone. This is the model legislation must follow. Public safety cannot depend on corporate goodwill. It must be mandated, resourced, and enforceable.

THE MUSIC INDUSTRY IS THE CANARY IN THE COAL MINE

If lawmakers want proof that unregulated AI already harms real people, they need only look at the music industry. It is the first cultural sector where AI extraction has fully industrialized.

Platforms like Suno, Udio, and LANDR show what happens when generative AI moves faster than law.

WHY MUSIC WAS HIT FIRST

Music became AI’s first cultural target because it is the predictable continuation of a century-long pattern. The music industry has always been the testing ground for technological extraction and artist exploitation.

Music was fully digitized first, has the longest history of systemic artist exploitation, and operates through centralized gatekeepers that prioritize profit over creative rights.

From payola to 360 deals, from Napster to streaming, each technological shift extracted more value while reducing artist control. AI represents the final frontier of this pattern: systems that can replicate not just distribution models, but creative labor itself.

SUNO AND UDIO: GENERATING WITHOUT CONSENT

Suno and Udio promise instant music generation including vocals, lyrics, production styles, and emotional tone. What they do not acknowledge is the invisible labor behind those outputs.

These systems were trained on decades of recorded music built by artists and communities, often Black, queer, underground, and global movements. That material was absorbed without consent, compensation, or attribution.

In June 2024, Sony Music Entertainment, Universal Music Group, and Warner Music Group filed copyright infringement lawsuits against both companies. The complaints alleged that the platforms unlawfully reproduced copyrighted recordings to train AI systems that directly compete with and devalue human artists.

The lawsuits cited examples where users replicated songs with vocals indistinguishable from well-known performers. The companies were accused of being intentionally evasive about their training data, suggesting disclosure would expose massive infringement.

Public availability does not imply consent for commercial replication. A book in a library may be read, not reprinted. That distinction matters legally and ethically.

In late 2025, Universal Music Group and Warner Music Group settled their cases and announced opt-in licensing arrangements. Sony Music Entertainment continues its litigation, and independent artists have filed class-action lawsuits alleging large-scale theft of creative work.

Calling this training data is linguistic camouflage. Without regulation, style becomes a legal loophole that allows creativity to be reproduced while its sources are erased.

LANDR: FROM CREATIVE TOOL TO CULTURAL GATEKEEPER

LANDR began as an assistive tool for mastering and distribution. Today it sits at the edge of a structural shift: platforms that both host creative work and train AI systems on user uploads.

In 2024, LANDR launched a Fair Trade AI program allowing musicians to opt their tracks into training datasets in exchange for revenue participation.

Yet a 2025 LANDR study found that most producers now use AI tools, many already use song generators, and nearly all plan to increase usage. When a single platform controls infrastructure and training pipelines, it gains power over aesthetic norms, market trends, and what sounds are economically viable.

This is private governance over culture without public oversight.

STREAMING DATA AND CULTURAL FLATTENING

Streaming platforms already map listener behavior and engagement. That data now feeds systems capable of generating music optimized for retention, not artistic intent.

AI-generated tracks are cheaper, faster, and infinitely scalable. Without regulation, platforms will prioritize them for economic efficiency rather than cultural value.

In 2025, Deezer revealed it receives over 50,000 fully AI-generated tracks per day, accounting for roughly a third of all uploads. A Deezer and Ipsos study found that nearly all listeners cannot distinguish AI-generated music from human-created tracks.

Investigations have already identified AI-generated artists appearing in algorithmic playlists and reaching hundreds of thousands of listeners.

This is market logic consuming culture.

WHY THIS IS AN AI SAFETY ISSUE

Systems without guardrails do not just create risk. They embed it.

Music shows how quickly this happens.

No consent
No dataset transparency
No compensation
No attribution

In 2025, a U.S. federal court ruled that training AI on copyrighted material without permission was not fair use, emphasizing the commercial and market-substituting nature of the practice.

Once these systems scale, they are nearly impossible to unwind. What begins in music will extend to writing, journalism, visual art, education, and identity itself.

Culture is not public data. It is a public good produced by people.

THE REGULATORY LANDSCAPE

Some regulation is beginning to emerge.

At the federal level, proposals like the CREATE AI Act aim to democratize access to AI research infrastructure. The TRAIN Act would allow copyright holders to investigate whether their works were used in AI training. The NO FAKES Act would protect voices and likenesses from unauthorized AI replication.

At the state level, California passed SB 53, the first law regulating frontier AI systems based on scale and risk. Tennessee passed the ELVIS Act, protecting musicians from AI voice cloning. New York enacted laws regulating synthetic performers.

Internationally, the European Union AI Act establishes binding requirements for high-risk AI systems and mandates copyright compliance and training transparency.

LAWMAKERS: THIS IS THE TEST CASE

Using music as a blueprint, AI legislation must include:

  • Explicit consent for training on creative works

  • Full disclosure of training datasets

  • Compensation for cultural labor

  • Restrictions on automatic training from user uploads

  • Mandatory labeling of AI-generated works

  • Protections against algorithmic displacement

These are not radical demands. They are baseline safeguards.

We already have the research.
We already have the evidence.
What we lack is law.

UNMIXED’S POSITION

We are not anti-AI.
We are anti-erasure.

When creative labor becomes data and artistry becomes content, the loss is not just economic. It is existential.

Music makes the danger audible.
Policy must make it stoppable.

Nina Katashvili
Previous

The Era of Erasure
Next

No Thanks, Just Truth: Nightlife, Land Back, and Late-Night Solidarity