Susan Harkins

Susan Harkins Logo

The machine learning art process, and the creative potential of generative AI.

AI Art is a term used to describe any type of artwork that has been created with the help of artificial intelligence. Traditional art methods involve a user utilizing either physical or digital tools such as brushes or software guided by the user’s hand to create artwork while AI uses advanced algorithms which generate visuals and patterns based on trained data sets and user input. Generative AI algorithms are trained using a process called machine learning. In simple terms, these algorithms are fed a large amount of data, such as images or text, and they learn from this data to create new content. Let’s focus specifically on how generative AI algorithms are trained for image-based AI tools.

The training process starts with a dataset of images that are sourced and used to teach the algorithm. These images can be of various types, styles, or themes, depending on the desired output of the AI tool. For example, if the AI tool is intended to generate landscape paintings, the dataset would obviously benefit from including a wide range of landscape images and paintings of various mediums. The algorithm analyzes the dataset and learns the patterns, styles, and features present in the images. It tries to understand what makes a landscape image look like a landscape, such as the colors, shapes, textures, and composition. Through this analysis, the algorithm develops a model that captures the essence of the dataset.

Once the training is complete, the generative AI algorithm can be used to create new images based on user input. AI Art technology can be used to create stunning visuals and captivating images, rendered in specific styles which are often difficult to achieve using traditional methods. When a user submits a text prompt into a generative image-based AI tool, the algorithm uses the knowledge it gained during training to generate an image that matches the input. For example, if a user types “sunset over the mountains,” the algorithm will generate an image that represents a sunset over mountains based on the patterns it learned from the training dataset. It will try to replicate the colors, shapes, and overall style of a sunset over mountains.

The output image is generated by the algorithm using its internal model, which is essentially a set of rules and patterns it learned from the training data. The more diverse and representative the training dataset is, the better the algorithm becomes at generating complex, realistic and visually appealing images, audio, and more. While explanations of generative AI models will stress that the output image is not a “replica” of the sourced data sets – it seems to be a Frankenstein monster with too many sewn together pieces for the human eye to perceive their origin.

Dana Rao Executive Vice President, General Counsel, and Chief Trust Officer at Adobe said in a statement “Generative AI also opens the door to new questions about ethics and responsibility in the digital age. As Adobe and others harness the power of this cutting-edge technology, we must come together across industries to develop, implement and respect a set of guardrails that will guide its responsible development and use.”

AI Art is rapidly growing in popularity due to its ability to generate specific, highly detailed, unique visuals in a relatively short amount of time with relatively little user input. If that sounds both scary and fascinating to you – good, it should. With advances in technology, AI Art will continue to evolve and become an increasingly important part of the art world. Like all great advancements in technology, how we use it and regulate its fair use will dictate if this tool is ultimately good or evil. Generative AI artwork can be leveraged for a variety of artistic purposes such as photoshopping images, creating new concept layouts, making music, rendering 3D modeling textures and more. It has even already been used to create an AI Supermodel, as if unrealistic body image standards weren’t already an issue – even when in the market of the human image, AI seeks to replace what is human.

Girl with a Pearl Earring - AI Background Fill GIF By Cinecom.Net on GIPHY

Can Machines Create Masterpieces?

Art has long been a platform for marginalized groups to express themselves, find solace, and share their experiences. It is a realm where human emotions and narratives are explored and conveyed. However, the rise of AI in art raises concerns about the removal of artists from the creative process. If machines lack the ability to feel, can they truly infuse artwork with genuine emotion? Art and masterpieces aim to connect the creator with the viewer, providing insights into the artist’s thoughts, struggles, successes, and emotions. A true masterpiece transcends visual perception, touching the heart and soul of the audience.

Currently, under most legal systems, machines or AI systems are not recognized as legal entities with the ability to claim ownership, creatorship, or copyright over the art or images they create. This is primarily because these systems are considered tools or instruments used by human creators rather than autonomous beings with legal rights. Ownership and creatorship are typically attributed to natural persons or legal entities such as individuals, companies, or organizations. These entities have the legal capacity to possess and exercise intellectual property rights, including copyright, which grants exclusive rights to reproduce, distribute, and publicly display their creative works.

Moreover, copyright law often requires a work to possess originality, creativity, and a human element to be eligible for protection. As AI systems operate based on algorithms and data inputs, they are generally seen as tools or mechanisms for executing predetermined instructions rather than exhibiting human-like creativity or originality. However, it’s important to note that legal systems and intellectual property laws vary across jurisdictions. There are currently many discussions and ongoing debates about potential legal frameworks to address AI-generated works and their ownership.

But what if the goal is not to create a masterpiece? What if the purpose is simply to design a logo or create a casual social media avatar? While AI art generators may seem appealing for such tasks, there are critical considerations for those intending to use generated content for commercial purposes.

Misconceptions surrounding AI art generators abound in the public consciousness. Many believe that AI generators create artwork solely based on user input, that the algorithms are trained on legally obtained materials, and that the AI is free from bias. In reality, these assumptions are often far from the truth.

When using an AI art generator for commercial applications, such as logos or products for sale, it is crucial to be cautious. The origins of the training data and the potential biases encoded within the algorithms are significant concerns. Unregulated AI components in commercial art can inadvertently perpetuate harmful stereotypes, infringe upon intellectual property rights, or even violate ethical boundaries.

Protecting the human element in art creation becomes paramount as AI continues to advance. Embracing technology as a tool rather than a replacement is essential. The human touch, creativity, and subjective experiences cannot be replicated by machines. While AI can offer novel possibilities, it should be utilized responsibly, with a deep respect for artistic integrity, diversity, and ethical considerations.

As the future of AI in art unfolds, it is crucial to foster ongoing discussions, promote transparency, and establish frameworks that protect the rights and preserve the authenticity of human artists. By balancing the potential of AI with the safeguarding of human creativity, we can ensure that technology enhances, rather than undermines, the richness and depth of artistic expression.

Ensuring your brand has an authentic voice that’s aligned with your mission goals and ethics goes a long way in establishing credibility and trust with customers.

Learn how to develop an authentic brand voice.

Click Here

AI-Generated Images and their Impact on the Digital Creative Process

Artificial intelligence generated art and generative design tools

Artificial Intelligence (AI) has undoubtedly made significant strides in the creative industry, offering tools that can generate art, music, and other forms of media. Graphic artists can effortlessly produce intricate images, and musical artists can compose complex melodies with relative ease. The advent of AI has also opened up new avenues for exploring creative ideas and pushing the boundaries of artistic expression. However, it is crucial to approach the integration of AI in art with caution and protect the essential human element in the creative process.

While AI-generated artwork may appear impressive at first glance, it is important to acknowledge that the results often require substantial editing and refinement by human artists. AI tools, despite their capabilities, still lack the profound understanding of emotion and context that humans possess. This can result in a lack of genuine creativity and originality in the works they produce. Human artists draw inspiration from various sources, and they are often transparent about their influences, which fosters a culture of attribution and respect within the artistic community. On the other hand, AI-generated artwork fails to inform viewers about the data sources that influenced its creation. It raises questions about the authenticity and true origin of the artwork, undermining the notion of creativity itself.

Furthermore, AI tools have limitations when it comes to thinking outside the box and producing truly unique and groundbreaking creations. They heavily rely on the data they are fed and, if not carefully curated, can perpetuate biases and reinforce existing imbalances in society. Without equal representation and inclusion of marginalized communities in the data used to train AI models, the resulting artwork may further marginalize these voices and perpetuate a lack of representation in the creative landscape.

The use of AI in the creative process should be approached with a critical eye, balancing the potential benefits with the need to preserve the artistic integrity and human touch. It is crucial to strike a balance where AI serves as a tool to enhance human creativity rather than overshadow or replace it entirely. As the creative industry embraces AI, it must also grapple with the ethical implications and actively work towards ensuring fair representation, transparency, and the preservation of the authentic human element that makes art truly meaningful.

Artificial Intelligence No GIF By ADWEEK on GIPHY

Legal and Ethical Concerns and the Imitation of Artists’ Works

AI image generators, artificial intelligence ethics, copyrights and AI image scraping

In order to create amazing works of art seeming from thin air, from a text entry no longer than the average tweet, there has to be a trade off, and there is, a huge one for artists and intellectual property law. The AI art generation software (generative AI) uses what data scientists call a neural network – which in the simplest terms is a complex set of machine learning algorithms which are used to identify patterns within immense quantities of data and provide some type of data output. The neural networks used to generate AI artwork are so complex that data scientists actually categorize it as deep learning. What this typically means is that the neural network has expanded exponentially through multiple layers of data revisions. In the case of most AI art generation systems the neural networks have already scraped the internet of various forms of data in order to classify what has already been created into these data pattern sets, in many cases this was done unethically under the guise of “educational” licensing.

Since these deep learning neural networks are programmed by humans, and fed sets of data, there is a pattern of inherent bias in AI technology. Perhaps the most prevalent example of this is in AI facial recognition software. A federal study by the National Institute of Standards and Technology released in December of 2019 concluded that most commercial facial-recognition systems are prone to bias and often not a singular one. Biases noted in the study ranged from issues identifying people of an older age, women and anyone not caucasian. How does this impact AI art generators? The same data issues exist regardless of the AI application. AI art can also be bias and create content from a narrow point of view than most would expect.

“Generative AI, as with any AI, is only as good as the data on which it’s trained. Mitigating harmful outputs starts with building and training on safe and inclusive datasets.” Dana Rao, Executive Vice President, General Counsel, and Chief Trust Officer at Adobe said this in a recent Adobe Blog from March 2023 titled Responsible innovation in the age of generative AI in which they addressed some major ethics issues plaguing the AI industry and what Adobe has been doing in order to help regulate trust. “Generative AI raises concerns over its ability to conjure up convincing synthetic content in a digital world already flooded with misinformation. As the amount of AI-generated content grows, it will be increasingly important to provide people with a way to deliver a message and authenticate that it is true.”

Just as counterfeiting a luxury handbag from a renowned designer is illegal, the use of generative AI presents a similar concern. This technology allows users to manipulate their own images to bear the likeness of celebrities or well-known public figures, often without obtaining their explicit permission. The potential negative legal consequences of this illegal use of someone’s likeness should not be underestimated, as there are several prominent examples that have set precedents regardless of the tools employed to create such deception.

One notable case that highlights the severity of such infringements is the legal battle between singer Ariana Grande and the fashion retailer Forever 21. In 2019, Ariana Grande filed a lawsuit against Forever 21 seeking $10 million in damages. The lawsuit alleged that after a failed endorsement pitch to Ariana, Forever 21 intentionally sought to exploit her trademark looks by using a look-alike model in their marketing materials. Moreover, there were allegations that the brand also used images of Ariana herself to promote its fashion and beauty products, without obtaining the necessary rights or permissions. This case illustrates the potential harm caused by the unauthorized use of a celebrity’s likeness. By capitalizing on Ariana Grande’s recognizable image without her consent, Forever 21 not only jeopardized her brand reputation but also likely increased their campaign profits by associating their products with her fame and popularity.

The lawsuit is just one example which emphasizes the importance of protecting the rights of celebrities and public figures, as their likeness holds significant commercial value. Such infringements not only violate the rights of the individuals involved but also have the potential to cause reputational harm, financial loss, and a breach of trust with fans and consumers. Protecting the rights of celebrities and public figures is crucial for maintaining the integrity of their brand and ensuring fair compensation for their image and likeness.

The Graphic Artists Guild published their initial concerns back in 2022 amidst the rise in prevalence of AI art and image generators and perhaps partially in response to the growing concern over intellectual property rights violations from artists within the community. In one post from January titled “AI Art Generators: Where We Stand” The Graphic Artist Guild sums up the current issues for the average designer or company looking to use these advanced AI Art Generators “The machine learning technology behind AI generators has raised serious questions about authorship, copyrightability, and ethics. As has often been the case when technology has disrupted traditional licensing and business models, the law provides few clear answers. The lawyers, policy makers, and enforcers behind our copyright system —  the engine which drives our creative economy —  are playing catch up.” For this reason it may be best for the average company or individual doing their due diligence to avoid the use of these specific generative AI models based on the Stable Diffusion Model, as the potential unknown liability is so unclear and much needed regulations are not in place to protect consumer or artist liability.

These are very well-known issues within the AI industry. Many companies, including Google, Apple and Adobe are continually researching how they can better integrate AI into their creative processes to ensure that the technology has a positive impact on content creators and the general population. While many other AI Art generation technology companies are currently flying by the seat of their pants attempting to rake in hundreds of millions of consumer and advertiser dollars before the law can catch up.

In another post from October the Guild announced that “The Graphic Artists Guild has issued a statement of concern to the IFRRO General Assembly on AI image generators, and the detrimental impact this technology is having on graphic artists

In this post they briefly outline the major concerns and also make one important note “There Can Be a Better Path Forward […]we believe that this is an opportunity for the tech sector to take steps to recognize the damage this technology can do to the livelihood of visual artists and to propose standards and take measures to prevent the abuse of creators’ works” Part of this includes establishing a way to “Permit visual artists to block the use of their works in the training of the AI image generator neural networks — for example, by leveraging image metadata” – something that seems to be the bare minimum of what would be ethically expected of any company – especially one using the works of others to build their for-profit product.

the simpsons GIF on GIPHY

The Better Path Forward

The future of AI image generators and artificial intelligence in visual arts

The future of AI image generators and artificial intelligence in the visual arts has sparked a profound debate surrounding the legal and ethical implications of content provenance. As AI continues to evolve and play an increasingly complex role in the modern creative process, concerns arise regarding the authenticity and ownership of digital artwork. In response to these challenges, the Content Authenticity Initiative has introduced the C2PA Guidelines, which propose safeguards to address these issues.

The C2PA Guidelines aim to establish a standardized framework for ensuring content provenance in digital creations. By leveraging blockchain technology, these guidelines provide a transparent and immutable record of an artwork’s origin, authorship, and alterations throughout its existence. This blockchain-based approach grants artists greater control over their work, enabling them to protect their intellectual property rights and maintain the integrity of their creations. Moreover, it empowers viewers and collectors to verify the authenticity and legality of an artwork, instilling trust in the digital art market.

While the C2PA Guidelines offer a promising solution, they are not the only initiative dedicated to safeguarding the rights of artists. The Human Artistry Campaign has emerged as a powerful advocate for the protection of intellectual property and the legal use of likeness. By promoting awareness and education on copyright laws and fair use, the campaign strives to ensure that artists receive appropriate recognition and compensation for their work.

In the face of the growing prevalence of AI-generated content, the Human Artistry Campaign emphasizes the value of human creativity and the distinctiveness it brings to the artistic landscape. By fostering a culture that values originality, the campaign seeks to curb the imitation of artists’ works by AI systems. It encourages artists, consumers, and technology developers to collaborate in creating an environment where ethical considerations are at the forefront of the creative process.

As we navigate the complex intersection of AI, content provenance, and intellectual property rights, it is essential to acknowledge the transformative potential of technology while upholding the principles that underpin the artistic world. By embracing the safeguards proposed by the C2PA Guidelines and supporting initiatives like the Human Artistry Campaign, we can forge a better path forward—one that champions both innovation and the protection of artists’ rights. Through these combined efforts, we can ensure that AI remains a powerful tool that augments human creativity rather than undermining it, preserving the richness and diversity of the visual arts for generations to come.