Skip to main content

   WEBSITE DESIGN   〰️   DIGITAL MARKETING   〰️   SOCIAL MEDIA MARKETING   〰️   WEB DEVELOPMENT   〰️  ARTIFICIAL INTELLIGENCE   〰️   GRAPHIC DESIGN  〰️  BRANDING   〰️

   WEBSITE DESIGN   〰️   DIGITAL MARKETING   〰️   SOCIAL MEDIA MARKETING   〰️   WEB DEVELOPMENT   〰️  ARTIFICIAL INTELLIGENCE   〰️   GRAPHIC DESIGN  〰️  BRANDING   〰️

   WEBSITE DESIGN   〰️   DIGITAL MARKETING   〰️   SOCIAL MEDIA MARKETING   〰️   WEB DEVELOPMENT   〰️  ARTIFICIAL INTELLIGENCE   〰️   GRAPHIC DESIGN  〰️  BRANDING   〰️

   WEBSITE DESIGN   〰️   DIGITAL MARKETING   〰️   SOCIAL MEDIA MARKETING   〰️   WEB DEVELOPMENT   〰️  ARTIFICIAL INTELLIGENCE   〰️   GRAPHIC DESIGN  〰️  BRANDING   〰️

Back in the mid to late 80’s, De La Soul and The Beastie Boys were chopping up vinyl and crafting hip hop’s iconic sound by sampling some of the worlds best known songs. It’s fair to say copyright wasn’t the first thing on their minds. The early hip hop scene was buzzing with creativity, but the legalities around sampling remained murky. Fast forward to today, and there’s a new frontier that’s making the waters of copyright just as unclear: Artificial Intelligence.

Sampling Data: The AI Connection

It’s no secret that AI has been revolutionising various industries, but its relationship with copyright is a complex one. Just as hip hop artists used samples to create something entirely new, AI models are trained on vast datasets, many of which include copyrighted materials.

From text-to-image models like Stable Diffusion to large language models like ChatGPT, the potential of AI is breathtaking. Yet, the questions remain: Is it infringement when the AI is trained on these copyrighted works? Can the creator be held liable? These are questions that remain unresolved, with pending US lawsuits still challenging the murky grounds (as of August 2023).

Training AI

AI models require massive datasets scraped from the Internet, containing images, text, music, and more. The catch? Much of this data is copyrighted. The situation has been likened to Google Books scanning copyrighted books, which was found to be fair use. But with new lawsuits in the AI domain, the lines are getting blurrier.

Some argue that the vastness of data used makes licensing all of it impractical, and restrictions would shift power to big corporations that can afford it. On the other side, regions like the United Kingdom, Germany, Japan, and the EU have incorporated exceptions for “text and data mining” (TDM) to their copyright laws, each with different conditions.

Copyrighting AI-Generated Works

Remember that “selfie” taken by a monkey? It’s not eligible for copyright protection, and neither are works created solely by machines in most jurisdictions. But what if a human exercises creativity in using an AI model? Could they claim copyright?

This opens a philosophical debate. If a photographer clicking a shutter button can claim originality, why not an artist who goes through several revisions to refine an AI’s output? The US Copyright Office has its own thoughts, emphasizing the need for the “manifestation of the author’s own creative conception.”

In the UK, a computer-generated work’s author is considered the person who made the “arrangements necessary for the creation of the work.” But who does that refer to: the programmer or the operator?

The Hip Hop of the Digital Era: AI in Music

Now, let’s hit a note closer to our hip hop analogy: AI’s impact on music. The emergence of platforms such as voicify.ai demonstrates how easy it is to mimic artists’ voices. Imagine creating a new song that sounds exactly like John Lennon singing. It’s been done already!

Let’s not even get into how The Beatles are releasing a ‘new’ song this year – 53 years after they broke up.

Cloning voices and song styling is reminiscent of sampling in the early days of hip hop. The creative process might be the same, but the legal aspects? Those are still up in the air.

Of course, there are flip sides to this for the artists as well. Grimes is offering her voice up for producers to use, at a cost, and artists such as Paul McCartney could use AI to de-age their voice to something more reminiscent of their early days. The possibilities are endless.

AI, Art, and the Copyright Conundrum

The lines between homage and plagiarism are more blurred than ever. Enter Jason Allen’s “Théâtre D’opéra Spatial,” a groundbreaking piece of art created not by brushes and chisels, but algorithms and data. More specifically, Midjourney.

When Allen set out to create this piece, he had something truly ambitious in mind. He asked his AI to craft an image in the styles of Leonardo da Vinci, the Renaissance master, and Alex Grey, the contemporary American psychedelic artist. The AI complied, weaving together a visual experience that resonated with both the past and the present.

But what does this mean for copyright law? Can you really copyright a style? And where does homage end and infringement begin?

The creation of “Théâtre D’opéra Spatial” opened up a Pandora’s box of legal and ethical questions. By tapping into the distinctive styles of renowned artists, Allen’s AI-powered creation walks a tightrope between inspiration and imitation.

The kicker? Jason’s AI-generated work of art won the Colorado State Fair’s art competition.

For Jason Allen to create his award-winning art, Midjourney was trained on over 100 million other pieces of art. That’s a lot of data.

If Jason, or Midjourney, infringed copyright how can this be addressed using existing laws? Is it a new form of fair us that permits use if the work is sufficiently transformed into something new?

Lawsuits and Litigations

From GitHub Copilot’s code editing tool to Stability AI’s training data issues, lawsuits are popping up left and right, putting AI and copyright under the spotlight. Even the technology’s occasional tendency to produce outputs identical to training images or to mimic an artist’s unique style has led to legal challenges. This is a can of worms that will keep giving for years to come……

Conclusion: A Long Loop…..

The early hip hop scene’s sampling practices opened up a world of creativity, but it also introduced complex legal questions. Today, AI’s use of data and its ability to replicate voices and styles echoes those same legal challenges.

Whether it’s De La Soul or a new AI-generated novel, the question of ownership and copyright continues to resonate and will challenge our understanding of creativity and ownership.

In the end, it may be the courts that decide the rhythm of this new era, setting the stage for the next wave of innovators, just as they did in the golden days of hip hop. Whether AI is a complex collage tool or a revolutionary new instrument, the legal needle has just hit the record.

PS. If you want to learn more about De La Souls magnificent debut album ‘3 Feet High and Rising’, and the history of the legal issues around it, check out the below video. It’s good lunch time viewing 🙂