AI In The Creative Industry - Trends, Thoughts, Surveys

The introduction of AI tools marked a significant paradigm shift for creatives across numerous fields. Although artificial intelligence has been around for a while, the broader understanding of its capabilities drastically changed after ChatGPT became popular among the public. Suddenly, and seemingly out of nowhere, new tools began to emerge, ranging from writing and design tools to AI for music production.

Fast-forward to 2024, and artificial intelligence has become somewhat normalized. Yet, among creatives, such developments are met with much criticism. According to a recently published Tracklib survey, 75% of music producers reject AI tools despite their increasing accessibility and popularity. Among those, 82% reject AI for artistic reasons, 32% say they are dissatisfied with the quality, and 14% do so for financial reasons. Interestingly, only 10% do so out of copyright-related concerns. Among the 25% who work with AI, 74% use it for stem separation, 46% for audio engineering-related tasks such as mastering, and 21% let AI generate elements for their songs. Only 3% use AI tools to create entire songs, which confirms that most artists continue to value human creativity.

Another study by Soundplate found that 80% of artists have never used any AI-powered music production tools, thus further emphasizing the preference for the more time-consuming, human-first approach. Finally, the Tracklib report shows that generative AI has a significantly worse reputation among artists than assistive AI, which is viewed somewhat more positively.

Such findings show that, despite the increasing accessibility, views on AI remain rather critical. The music industry’s concerns revolve around copyright infringement and voice cloning through deepfakes. Others worry about their job security, a topic widely discussed among producers and audio engineers concerned about partial or total replacement.

However, a convincing AI-generated product requires good prompt writing, and sometimes, even a solid prompt does not help create a high-quality beat, track, or artwork. And artists agree. Not too long ago, WIRED asked professional musicians to test a new, ‘ethical’ AI music generator trained on licensed material. The results? All artists found the outcomes rather uninspiring. A key criticism was that they sounded too clean, too polished, hi-fi, and… too mechanical. Others thought the results sounded somewhat cliché as if the AI tried to play it safe.

Still, one artist explained that despite its current limitations, AI has the potential to replace numerous jobs. While complex compositions will still require human creativity, simple works can already be generated faster and at lower costs. This situation is especially challenging for creatives who work for companies rather than going down the fine arts path, as they depend on clients who often prioritize financial matters.

So, how much should we worry about AI? To better understand the topic, it may be helpful to look at the development of technology over time.

Throughout history, technology has constantly evolved, disrupted, replaced, and created entirely new fields. Let's look into some concrete examples. In the past, painters were needed to accurately depict people and places. With the introduction of analog cameras, many painters were no longer the primary choice for such tasks. Soon after, the appearance of digital cameras replaced many jobs of those educated in analog photography, which requires numerous skills for taking and developing pictures. The introduction of high-quality phone cameras rendered photography significantly more accessible to all, and while it did not replace professional photography, it opened new doors for individuals to produce visual content on their own.

Regarding music production, the transition from analog to digital to virtual has brought about significant changes. Previously, artists required access to expensive gear that took up a lot of space. Today, the majority of work has shifted into the virtual realm, with virtualized tools being relatively affordable compared to analog ones. Similarly, today's DSP-based VSTs allow anyone to add different types of instruments to their tracks, while plugins can help situate the sound in a wide variety of spaces, such as churches or hallways.

These developments have their downsides, as they can diminish the number of jobs previously in high demand. At times, they can even eradicate entire fields. However, they have also rendered numerous art forms more accessible, whether through simplification or affordability. Today's music producers do not necessarily need much more than a computer, music software, and plugins. Those aiming to dabble in photography can start with a regular phone camera and basic photo editing skills. Aspiring designers, too, only need a computer, software, and digital assets. These benefits apply to professionals and hobbyists alike, especially those just starting out and testing the waters.

Returning to AI, it is crucial to understand the distinction between assistive and generative tools before further discussing the dangers and benefits. This is because assistive tools enhance existing capabilities, while generative tools create things from scratch.

Consider the scenario of a graphic designer using AI to quickly cut out an object from an image, thereby saving a significant amount of time. Such tools fall into the category of assistive AI and are not necessarily detrimental, as they can streamline and simplify technical processes without infringing on creativity. Using assistive tools for technical yet non-creative tasks can be particularly beneficial for professional or freelance creatives who frequently face tight deadlines and juggle multiple projects simultaneously. They can also be helpful for hobbyists who have limited time for their creative pursuits due to their day jobs.

That said, it is important to recognize that assistive tools also have limitations. For instance, AI sample matching can save much time producers would otherwise spend on endless scrolling. But it can also infringe on random discoveries lying at the core of many iconic tracks and even entire subgenres. Thus, while assistive AI can help in technical contexts, it can lead to standardization if experimentation is replaced with convenience and quantity over quality.

Concerns are more prevalent regarding generative AI due to its disruptive nature and capabilities to create entire works from scratch. Those interested in the topic know that AI is becoming increasingly good at generating realistic images, videos, or 3D animations in various styles. It is also powerful enough to realistically clone voices and generate entire instrumentals (see video below). For this reason, it is valid to view generative AI as a threat to the creative industry, especially when it comes to entry-level jobs and simple work that previously required human skills.

Nevertheless, generative AI does have a few positive sides. For instance, some people may have great creative ideas but lack the time to develop the necessary skills to turn them into finished products. Others may benefit from generative tools to get inspired, and some may use them to create elements or field recordings that they cannot find anywhere else. In the latter two cases, generative AI can be quite helpful and used in a rather assistive manner without taking away from the creativity of the human mind.