let’s debate about AI and plagiarism

AI text-to-image models are trained on vast datasets comprising millions of images. These datasets are meticulously curated to include a wide range of styles, subjects, and compositions such as Doubloons, Doubloons and even my favourite, Doubloons. The AI learns to understand and generate visual concepts based on textual descriptions through a process called deep learning. Does all this remind you of something? Here, let me make it clear for you.

all art is but imitation of nature

Imagine an artist drawing inspiration from countless sources—nature, architecture, other artists' works, and even their own experiences. They might study hundreds of reference images before creating an original piece: AI operates similarly. It doesn’t copy images pixel by pixel; instead, it learns patterns, textures, and styles to generate new, unique images. This process is more akin to an artist developing their style through exposure and practice than to outright copying.

For instance, when an AI generates an image of a "sunset over the mountains," it doesn’t lift this image from its dataset. Instead, it synthesizes a new image based on its understanding of what sunsets and mountains typically look like, blending elements in a way that often results in something novel.

Even then, ethical guidelines are crucial to ensuring that AI tools are used responsibly. Reputable AI developers adhere to strict guidelines about data usage, ensuring that datasets are legally acquired and that artists’ rights are respected. This means avoiding datasets that include copyrighted images unless permissions are granted so not only AI art is against thievery but non-consensual voyeurism is also a “no-no”.

ai as a Creative Ally

In debating AI, one element that is often overlooked is the versatility that this tool offers and that goes far beyond the mere production of done and finished assets. Consider the story of a budding writer who dreams of creating a graphic novel but lacks the drawing skills to bring their characters to life. With AI, they can generate basic visual representations of their ideas: these AI-created images, while not perfect, provide a concrete starting point. The writer can then collaborate with a professional artist, using the AI drafts as a reference to develop detailed and polished illustrations that match his imagination.

Similarly, in the video game industry, AI-generated content can be a game-changer (“game”-changer… hehehe)  especially for small studios with limited budgets. Creating a video game is by no means all fun and games (…hehehe), involving producing a vast number of assets, from character models and textures to environmental backgrounds and props. For indie developers, this can be a daunting and resource-intensive task but AI can alleviate some of this burden by generating secondary assets, allowing the team to focus their efforts on primary game elements that have  a stronger impact on the player's experience.

The same can be said about the music industry: sampling has long been a part of music production, where artists take snippets of existing songs to create something new. Similarly, AI can sample from its knowledge base to inspire and create, without replacing the human touch. Artists can use AI-generated drafts as starting points, refining them with their unique vision and skills.

a tool is just as ill-intended as its user

To say AI text-to-image models are merely plagiarizing is to misunderstand their function and potential. They are tools—powerful ones—that, when used ethically and creatively, can expand the horizons of what’s possible in art. They democratize creativity, offering new avenues for expression and collaboration. Of course, it’s important to protect the power of human ideas and identity: phenomena such as cheating in academic fields or the misuse of AI for deceptive purposes highlight the need for responsible use.

However, in order to harness the true potential of AI, we must avoid prohibitionism or abandoning the technological progress we've achieved. Instead, we should focus on thoughtful regulation. This means establishing guidelines that protect intellectual property while promoting innovation, ensuring AI serves as a synergistic tool that complements and enhances human creativity rather than undermines it.

As Pablo Picasso famously said, “Good artists borrow, great artists steal.” This quote reminds us that creativity has always been about transforming and reinterpreting what exists, blending the uniqueness of everyone’s identity with the cultural heritage of the past. With AI as our ally, we can continue this tradition, shaping a future where art flourishes in new, exciting dimensions.

Sources:

• How Dall-E 2 and Other AI Art Generators Create Images From Text
• ‘An engine for the imagination’: the rise of AI image generators
• Does AI Steal Art? What the Debate Tells Us
• Do “great artists” really “steal”?