1888 Advertisment for Kodak camera, source: THE GUARDIANTimeline by Fabian Mosele, see the finished interactive timeline HERE

When photography was invited in the early 1800s, it was primarily a tool for scientists to experiment with photo-sensitive materials. During this time, futurists imagined potential use cases of photography, and artists debated if photography might be considered art. Fine art societies worked to develop standards for how photographs should look.

[Photographs are useful so long as they are taken in] accordance, as far as it is possible, with the acknowledged principles of Fine Art.

Upon Photography in an Artistic View, 1853, Journal of the Photographic Society

A couple of decades later in 1888, Kodak launched a camera Kodak’s with the tagline “You press the button, we do the rest”. A century later, the button has been replaced by a cursor.

Image capture has always been important for society and people. Images help us document historic events, persuade others to join our causes, capture every and any moment of everyday life. Images have power to keep the memory of a loved one alive, evoke buried emotions, and help us tell our story.

Through the year, so many tools have revolutionized the field, from Camera Obscura to the Kodak Camera, the Polaroid Camera, and the iPhone. Now AI tools push us even further. The use of AI allows us to create memories of events that never happened – or even images of people who don’t exist.

If an image is worth a thousand words, how does AI change our stories and how we tell them?

📸 WIP Timeline by Fabian Mosele see the finished interactive timeline HERE

Listen to Fabian discuss how Artificial intelligence is changing our creative process, how it is connected to semiotics and what are artists currently doing with this one model – VQGAN+CLIP AI has been around for decades, but it’s only recently that it has become more accessible for artists and creatives; low-code applications and new educational resources launch every day.As a result, a new wave of digital art and artists have emerged. Generative Adversarial Networks (GANs) have launched its own distinct art style as seen in music videos by Cuco ABRA, and Magdalena Bay.

How has generative image-making tools impacted image-making? How will it affect the future of image-making? AIxD member Fabian Mosele, says:

It is making us think about our creative process in a different way – a way that is only possible because of these tools. Now is the time to dig deep and see what can be done.

FABIAN MOSELE, AI-PROMPT ENGINEER + 3D ANIMATOR , AIXD MEMBER 🤖

Bias

Artists are not only looking at AI to generate images, but also to address the impact of AI systems on society, the harms of algorithmic biases and its negative impact on social justice, equity, and inclusion. Many of these AI models are trained on images from across the interest, using the corresponding alt-text as its defacto image label. This leave these models vulnerable in two ways, as described below artist @kaliyuga_ai:

It is just an infinitely complex mirror held up to our society. It’s trained on images we deem worthy enough to post on the internet and the language we choose to describe them.
– KALIYUNG, ARTIST, SOURCE: THE AI THAT CREATES ANY PICTURE YOU WANT, EXPLAINED, VOX

‘ImageNet Roulette’ is a project created by artist Trevor Paglen and researcher Kate Crawford in response to their concerns about the systemic biases in ImageNet (a free repository containing over 14 million images). These images were manually labelled as part of a Stanford University project to “map out the entire world of objects” and are widely used by researchers to train AI systems.

There is no easy technical ‘fix’ by shifting demographics, deleting offensive terms, or seeking equal representation by skin tone. The whole endeavor of collecting images, categorizing them, and labeling them is itself a form of politics, filled with questions about who gets to decide what images mean and what kinds of social and political work those representations perform.

— KATE CRAWFORD AND TREVOR PAGLAN , EXCAVATING AI: THE POLITICS OF TRAINING SETS FOR MACHINE LEARNING (SEPT 2019)

But because the images were labelled by humans, many of the labels are subjective and reflect the biases and politics of the individuals who created them. In their research, they racial bias, gender bias, and offensive stereotypes within the image labels, from racist slurs to misogynistic terms. Paglen and Crawford’s project went viral, leading to more than half of the 1.2 million pictures in the dataset’s ‘people’ category being erased.

Deepfakes

Deepfakes are AI-generated images or audio that replace the likeness of the original person in the image with someone else. We see deepfakes being used creatively seen in Kendrick Lamar’s music video The Heart Part 5, or when the Mona Lisa was bought to life by Samsung AI. While he European Film Market Horizon classified AI as a non-threat to creativity, it See an example of this type of deepfake below, published by Bloomberg Quicktake. also warned about its potential for misuse.

Deepfakes aren’t well-regulated, leaving people who are victims of deepfakes without the meaningful avenues of redress. Additionally, deepfakes can fuel disinformation by creating videos of famous people or politicians saying things they never said.

OWNERSHIP & COPYRIGHT

While there’s very little legal protection, there are a lot of concerns around copyright and creative ownership in regards to generative images. David O’Reilly summarizes it well in his recent Instagram post (below):

We agree with Lenka that public literacy of data and AI is an important part in ensuring these new and powerful tools are developed for the many and not the few. Please join us as we continue to learn tinker, and play!

Here’s a list with some of our favourite resources:

• Artificial Images youtube channel
• Deep Dream Generator - Stylise images using enhanced versions of Google Deep Dream with the Deep Dream Generator.
• Google Deep Dream - GitHub repository for implementing Google Deep Dream.
• ArtBreeder – Merge images together to create new pictures, make hybrid AI portraits and create wild new forms that have never been seen before.

Additionally, we’re constantly curating resources & tools to help you explore how AI can augment your creative practice in our Resource Library.

One thing is certain: the future of imaging, photography and video will not be shaped solely by professionals working in those fields. Using the accessible tools mentioned before, anyone can experiment and create new artworks, whether they know how to code or not. As generative visuals are acknowledged as art, the border between fine arts and vernacular visual languages will become more and more blurred, making us rethink the definitions of art and visual culture.

Along these lines, there has been a rapid shift in how art is made and consumed with the emergence of technologies such as DALL-E and Midjourney. As we are able to quickly generate high-quality images from text prompts, our thinking is adapting to these new formats and our creativity is taking new forms.

For artists working in the field, these tools can become damaging to their industry on a larger scale. Some creatives have expressed their discouragement of the use of AI, considering the scenario where large corporations take advantage of the efficiency and quick turnaround of AI tools in comparison to commissioning artists. As art shouldn’t be a means to an end, we should strive to collectively build and use these tools as a way to empower creativity, instead of replacing the need for artists. Being an independent community of practice, we work to create a space for exploration, meaning that exploration consists both of joyful hands-on experiments, but also conversations and contemplating the infrastructure behind the tools we like to play with.