Advances in Generative AI Bring the ‘Creative Singularity,’ But Are We Ready?

Written by: Jesse Damiani

Source: FreeThink

Image source: Generated by Unbounded AI Gongju

In 2022, text-to-image generators like Midjourney, OpenAI’s DALL-E 2, and Stable Diffusion begin to enter the public eye, sparking a lively debate about the role of “generative AI” in creative production. And when ChatGPT launched in November and crossed the million-user mark within five days, concerns about AI job displacement expanded beyond visual artists to include writers, journalists and copywriters. Then there was a wave of new AI products from startups and incumbents alike, from voice cloning and avatar replacement to interface design and code cobots. Suddenly, it feels like no category of human creative labor is immune to automation. After months of growing outrage over their art being included in training data without permission or compensation, thousands of artists joined the “No AI Art” online protest in December.

Like chess before it, art is considered a quintessentially human achievement, but the rapid development of generative artificial intelligence is complicating this thinking. Algorithms have been everywhere for years, affecting how we work, socially and play, but the stream of innovations in generative AI has surprised even skeptics. Concerns about these technologies have also sparked renewed discussion of the Singularity, a term used to describe the moment when technological advancement, often in the form of “artificial superintelligence,” explodes faster than human capabilities, bringing unforeseen consequences to civilization. changes, and presents a new reality in which humans are replaced (think hostile robot takeover in science fiction).

I don’t think the “singularity” has arrived, and there are problems with building machine intelligence by generalizing the human perspective. However, I do believe that recent advances in generative AI have pulled us into a not-so-apocalyptic “creative singularity,” upending the fundamental norms of creative production and related industries.

To be honest, I’m a bit put off by the “disruption” talk, but after a decade of working in emerging technologies as a writer, curator, and futurist, I have to admit that it’s not like the usual business artists, scholars Mat Dryhurst Said recently: “People are understandably tired of those who work in technology saying that everything is going to change, and unfortunately, as it stands right now, that’s all going to change.” The emergence of large multimodal models, Things like Google’s PaLM-E and OpenAI’s GPT4 — the latter, which the researchers claim shows the “spark” of “artificial general intelligence” because of its ability to solve problems across domains without special prompting — feel like artificial intelligence Indicators that affect the future. At the same time, many enthusiastic proclamations by prominent figures in the field feel disconnected from reality. Moreover, they ignore less extreme but more applicable lessons from the past that must be reiterated in the face of hype that often erases instructive lessons.

What should we do if the creative singularity has already occurred? To see this, I draw on a framework I developed based on philosopher of science Thomas Kuhn’s concept of a paradigm: a set of concepts, theories, or patterns that form an explanatory model of global organization. This framework views reality as something that evolves with humans—it is the sum of our capacities to agree on what is real. The technologies and symbol systems we create (literacy, numeracy, codes, etc.) actually expand the content of reality. Large-scale machine learning tools are an integral part of a contemporary paradigm that I call “post-reality”. If, as Marshall-McLuhan put it, art serves as a “distant early warning system that can always tell the old culture what is going on with it,” then artists and other creative professionals who have been researching, experimenting, and using artificial intelligence Contributions provide key signals to this new world.

alien life form

In a 1999 interview, David Bowie tried to convince journalist Jeremy Paxman that the Internet was not an incremental innovation but a sea change in the way art was created, distributed and experienced. “I think we’re actually on the edge of something exciting and scary,” Bowie said. “It’s not just a tool, it’s an alien lifeform.”

Today, there is a similar sentiment among proponents of generative artificial intelligence. But what exactly are the new possibilities? Will new patterns of creativity and social dynamics emerge, and what are the implications and by-products of this change? Who will be most affected? To distinguish signal from noise, these questions will be analyzed from four perspectives, including automation and creative labor, intensification and speed, aesthetics and artistry, and fusion and emergence.

Automation and Creative Labor

On the surface, it’s easy to understand that billions of people now have access to powerful tools for creative expression, or that artists can automate tedious parts of their practice. But the elephant in the room—the one that most directly affects most people—is how these tools intersect with work and livelihoods.

Mashinka Firunts Hakopian, associate professor of technology and social justice at the ArtCenter School of Design and author of The Institute for Other Intelligences, has studied artificial intelligence for nearly a decade, focusing on the interaction of AI with real-world systems. This speculative fiction examines how the myths surrounding technology often obscure the realities of its creation. In it, she extends philosopher and historian of science Donna Haraway’s critique of so-called “objective” systems to artificial intelligence. She emphasized that all data showed an implicit political overtone, often reflecting existing power structures.

“The issues we’re seeing about workforces now are a continuum of issues that have arisen over the years around emerging technologies and workforces, and they reproduce many of the same gaps and omissions,” Hakopian said in an interview. “For example, the debate around generative adversarial networks and image generators has been based on the labor of the artists being extracted, but little attention has been paid to the labor of the data workers who are training these models, and the labor conditions within them. They are accepting training.”

The notion of creativity as we understand it is encapsulated in historical norms that influence the types of expression that are considered valuable (and thus reproducible).

“What forms of visuals and whose visions were highlighted, replicated, extracted, or remixed in the output we’re seeing now?” Harkopian said. “There’s a strange paradox that we think these tools are very novel, but oftentimes what they produce ends up being replicas of existing classics.”

In addition to using generative tools to create, artists also play an important role in reflecting on the absence of tools – “misusing” them in order to discover their weaknesses. Artist Minne Atairu uses machine intelligence in a range of different projects, including an examination of algorithmic aesthetic standards and a reimagining of Benin bronzes in her Lumen Award-winning series IGÚN. Works like this demonstrate how artists rigorously combine these tools to generate novel artistic expressions, even interrogating the biases of the models used to create their work. But Hakopian cautions that we’re skipping a critical step in our search for artists and designers who are successfully adapting to a fundamental shift in creative work.

“The burden of responsibility should instead be borne by the infrastructure layer of the tech companies that produce these technologies, the employers and clients who recruit labor in this art and design ecosystem, and regulators,” Hakopian said.

When it comes to the workforce, AI appears set to follow current economic and political models rather than replace them. Moreover, with large incumbents, start-ups, and government agencies locked into the AI ​​innovation arms race, it is easy to see how the multipolar trap is exacerbated by the creative singularity: To reduce costs, companies consolidate their workforces , offloading creative tasks to gig (or even “ghost”) work, and offsetting the growing number of creatives who then compete for dwindling job openings. In fact, this has already happened.

Of course, new jobs will also be created thanks to generative AI, and more nuanced than the headlined list of suggested engineers. Moreover, as large incumbents, start-ups, and government agencies are locked into the AI ​​innovation arms race, it is easy to see how the multipolar trap is exacerbated by the creative singularity: in the quest to reduce costs, companies consolidate their workforces , offloading creative tasks to temp (or even “ghost”) jobs and offsetting the growing number of creatives who then compete for a dwindling number of open positions. In fact, this is already happening.

Of course, generative AI will also create new jobs, and more nuanced than the headlined list of prompted engineers. But the open question is whether that number will keep up with the jobs being cut, and how the surrounding industry and government will respond to the shock.

Strengthen and speed

Generative AI augments our creative capabilities and the speed of content production. Alexander Reben is an artist and roboticist whose artistic research and experiments use humor and the absurd to reveal the potential and limitations of artificial intelligence. Over the course of our collaborations, including his solo exhibition at the Crocker Museum, AI Am I, and the forthcoming book I Create Like the Word: Poetry in the Age of Machine Intelligence, Reben and I have been discussing what he calls “the Human-machine symbiosis”, this is the research direction he has been pursuing since 2012. The term is a twist on the more traditional “human-robot collaboration,” not just a semantic flourish. It simultaneously reflects his belief in the role of technology in human evolution and positions his encounters with machines as expressions of emerging relationships with learning entities rather than mere inert art materials.

“The idea of ​​man-machine symbiosis comes from the fact that technology is something that is inseparable from humans,” Reben said in an interview. “Inventing stone tools and other external means of amplifying our capabilities allows us, for example, to have more calories and time to do things like invent science and philosophy. Technology has always been a very human thing.”

From stretched canvases to paints, all art-making tools were once new technology. Generative artificial intelligence is the latest in a long line of such innovations expanding our ability to create art. But what’s unique about these new tools—particularly new “AI agent” products like Auto-GPT—is the degree to which they are agents and self-learners in the creative process.

“The type of automation we’re seeing now is different than what we’ve seen in other periods of automation like the Industrial Revolution,” Reben said. “We’re automating mental and physical work right now, but I don’t think we’re fully prepared.”

Suddenly having the ability to augment our minds in this way will raise complex questions about the way viewers experience art and creative expression. Lauren Lee McCarthy is an artist and associate professor in the UCLA Design Media Arts program and creator of p5.JS, a JavaScript library for creative coding used by many Crypto artists around the world. As someone with hands-on experience developing Crypto art and tools, she wondered about the ripple effects of this new mode of cultural production. In her view, the speed of new AI tools will affect who the algorithms are “discoverable,” perhaps favoring those who can quickly produce and share content over those who spend more time developing their work, in her view. . In turn, this may affect the public’s ability to connect with meaningful art.

“I think we’re going to see fewer jobs like this because it’s less economically sustainable,” McCarthy said. “If this does become a trend, it will mean a huge loss of culture, and our ability to process and understand the world through the art culture around us. Because that’s what art does: it provides us with a way to understand what is going on. way. So everything happens faster and less work is created in the right time and space.”

In 1930, renowned economist John Maynard Keynes predicted that by the early 21st century, technological advances would usher in an “age of leisure and abundance,” with a 15-hour workweek. Would deploying generative agents on general-purpose tasks open up more free time for humans to more fully explore their own creativity? This is a vision that many will embrace. However, even in the most optimistic scenario, the road will be bumpy, as generative tools may also create new forms of distraction while increasing productivity, but one step in this direction is the collaborator (copilot), a lightweight AI agent that can play different roles. For creative tasks, robots act as “blank page killers,” helping both trained artists and novice creative endeavors jumpstart their creative endeavors, whether it’s helping visualize the opening of an essay, the concept of a character, or the beginning of a series of drawings. simulation model.

For non-artists, collaborators may automate work to free up time for creative tasks, either as personal assistants or agents for specific intellectual tasks. Just as there was once “an application”, we can now imagine “a collaborator”, although we have to account for illusions and potential alignment issues. This augmentation is a double-edged sword; for some it promotes unstructured time, for others it increases competition and demands on their time (and may ultimately divide people from The interaction between them differentiates into a “luxury” experience). But the tendency to offload intellectual tasks to collaborators can foster the deep value of more human capabilities: imagination, curiosity, synthesis, presence, and interconnectedness—at the same time that creative capabilities can be decentralized through generative tools to in people outside of creative occupations.

aesthetics and artistry

The creative singularity will mean a shift in aesthetics and the way artists work. McCarthy explained that generative AI creates a new social environment in which the public will look to artists for reflection, rather than worry that it will invalidate human efforts.

“I think the role of the artist is always to use the media tools available and offer a unique human or artistic perspective,” McCarthy said. “I’m not sure if this can be automated.”

Maybe ChatGPT will be the bot that pushes out a billion books, but how many will spark the interest of the public to actually read them? How many of these people would consider these books to be competitive with books written by people? In this sense, the highest and most conceptual forms of creation—derived from artists’ deep engagement with the world, their craft, and the questions that motivate their practice—seem insulated from automation. If anything, the public will need these interrogations more than ever to figure out what happened. For these artists, the generative engine will join a range of possible tools and materials that might be helpful in making a given artwork.

As AI capabilities empower more people to execute professional-quality creative output, they will continue a decades-old trend in art: focusing value on the combination of conceptual and aesthetic execution. A certain percentage of that number will be those who would otherwise feel barred from participating in art; if they have strong enough ideas (at least in theory), they will be able to produce meaningful conceptual art. On the other hand, more people will be able to engage in new creative activities for fun rather than to pursue a job or career.

“The greatest potential lies in this democratization of expression, where people are able to create output based on their imaginations, which may have been difficult for them in the past, whether due to lack of skills, capabilities or knowledge,” Reben said. “An obvious example is the camera. In the past, photography required high skill and, moreover, chemicals like cyanide to handle it. Now, everyone has a camera in their pocket.”

In the 2010s, art created with generative adversarial networks (GANs) and other forms of machine learning had a distinctive look that can be seen in the work of Memo Akten, Sofia Crespo, Jake Elwes, Mario Klingemann, Anna Ridler, and others. Ingrid Hoezl and Remi Marie called it “soft image” (and later “post image”), where the image-based work “was no longer a representation of a solid world, but… a programmable Database views. These aesthetics have given way to more plausible and realistic output (see: pope in a puffy coat). But even as generative tools produce increasingly human-like images, the creative singularity gives rise to new aesthetics Author and musician K Allado-McDowell identifies four “side effects” of using a text-image engine: hallucinations, hybridity, mutated language, and appropriation.

“Wet clay dictates potter’s movements; AI system shapes mind by subconsciously ingesting word/image maps,” McDowell writes in Side FX. “The inner world of the neural network is mined and imitated in the model of the artist’s inner world.”

These human-machine feedback loops created as a composite of human beings, however flawed and biased, represent a new historical context for creativity. The myth of the “lone genius” artist has long been maligned, and its relevance further diminished by the creative singularity. It also means that models will play a huge role in the aesthetics we encounter every day—whether it’s the way ChatGPT generates language, Midjourney generates images, or Runway inserts videos. Without input from a wide variety of players, this risks homogenizing creativity rather than amplifying or enhancing it. Furthermore, the limitations of models and datasets will also determine the visibility of a given medium. For example, art mediums that are easily packaged for machines—text, flat images, and sound—have so far generated more attention, investment, and innovation. Over time, this can affect who encounters different types of art, and the decisions artists make when choosing a form. For the general public, the proliferation of machine-generated content may even have a profound impact on their understanding of reality.

This also raises questions around the imitative modes of art making mobilized by generative AI: cover, knockoff, and imitation. After growing number of AI-generated songs, Grimes AnnounceRoyalties will be split equally with any songs cloned using AI voice.

Even beyond legal or technical considerations, the ability to imitate other artists has profound implications for how artists develop their craft. Steve Jobs famously quoted Faulkner (quoting Stravinsky): “Good artists copy, great artists steal.” An important aspect of artists developing their poetics, style, and unique language of expression The method is to delve into the works that inspired them. By taking influential works as sources, analyzing them, playing with elements, and reconfiguring them, artists come into their own. And that will change when we reach a point where anyone can generate a high-fidelity imitation at the blink of a cue.

Holly Herndon and Mat Dryhurst popularized the term “spawning,” the use of artificial intelligence to create works that resemble others. In response to “heart on my sleeve,” Herndon distinguishIn order to be able to imitate an artist and bring the same level of care and artistic intention. When everyone can cover in the style of every other artist, a collective culture emerges that shapes not only how young artists learn, but who they meet. Recent Wes Anderson trends on TikTok and Instagram also suggest that well-known artists may be influenced by AI-generated riffs — compressing their work into exaggerated or stereotypical representations.

Writer Ted Chiang argues that for young writers, the output of any kind of big language model is not a helpful starting point: “If you’re a writer, you’re going to write a lot of unoriginal stuff before you write original stuff. works. And the time and energy spent on these non-original works is not wasted….which is what enables you to create original works in the end. Spend on choosing the right words and rearranging sentences to better interact with each other Timing in articulation can teach you how to communicate meaning through prose.”

Imitation is thus another important (and tricky) feature of artist education and machine aesthetics, and the only thing we can be sure of is that this phenomenon will prompt a major shift in the way artists develop their skills.

Fusion and Emergence

The creative singularity also means that creative ability will become common among people who would not otherwise think of themselves as creative. Given existing precedents—think user-generated videos on YouTube—culture can change in unexpected ways and drive new mediums and forms of expression when capabilities become pervasive.

“Maybe our understanding of what constitutes creative output will change,” McCarthy said. “That might end up being different than someone generating an image or text or something, which might become more prevalent and used more in the way we communicate with memes today.”

Stephen Marche has dubbed the coming era the Big Blur, because all written content will be accompanied by the question: “Man or machine?” This blur, I believe, goes beyond the provenance or veracity of the content we encounter, the actual It has fundamentally changed the way knowledge is produced, organized and applied in post-reality. As the creative impulse (however historically conditioned) permeates other fields, it will cause a deeper shift: best practices and insights from other fields will feed into each other. Creativity becomes a hemoglobin that transports ideas across domains.

Economist Noah Smith has dubbed artificial intelligence “the third magic,” calling it a massive meta-innovation that follows developments in history (transmitting information) and science (deriving general principles about how the world works), updating the way we understand the world. One way artificial intelligence—particularly deep learning—differs from the scientific method is in its ability to identify patterns in vast amounts of data without any particular idea of ​​what it should find. This approach to obtaining information means that, in many cases, insights are available but not necessarily interpretable (through so-called “black box” questions).

“[M] Any complex phenomenon like language has underlying regularities that are hard to generalize but still generalizable,” Smith wrote. “If you have enough data, you can create a model (or, if you prefer, An ‘AI’) can encode many (all?) of the extremely complex rules of human language and apply them to conversations that never existed before. “

We have already described in the “AI cryptids” Crungus and Loab and the putative secret language of DALL-E 2 (i.e. Apoploe vesrreaitais) for a glimpse of these strange possibilities.

In this way, our intellectual paradigms become more aligned with the workings of creativity, following Alfred North Whitehead’s assertion that art “is the imposition of a pattern on experience, and our aesthetic enjoyment is the recognition of the pattern.” However, This “control without understanding, power without knowledge” relationship to knowledge requires strong safety, ethics, and “slow AI” devices in the public and private sectors to advocate for fair models and development processes that prevent the Exacerbation of biased outcomes witnessed by algorithmic culture (e.g. predictive regulation, loan evaluation). In addition, it asks individuals not to rely solely on the AI ​​products of big tech companies whose markets and stakeholder incentives may limit the forms of creativity that can be explored in the first place.

The Creative Singularity in Post-Reality

In an article written in the early days of Covid-19, Elizabeth Dias outlines how apocalypse, when understood through its original Greek usage (apokalypsis), means a revelation or revelation rather than the end of the world. Through this lens, the Singularity appears to be a key unveiling point on the continuum rather than a single, exaggerated cataclysm. As much as I get annoyed by dogmatic Singularism, I do believe that the Creative Singularity is a revealing moment, an important development in the evolution of human creativity. It forces us to confront the ways in which creative labor has been devalued long before the advent of generative tools, and how outside forces are using AI as an accelerator. It also reveals the value of human curiosity, critical thinking and analysis — which still cannot be easily automated, and which are critical to transforming what is happening in the “old culture”. Different people will feel differently about its effects. For some it will bring about a dramatic shift in day-to-day work, for others it will spark new creative tendencies that might otherwise have been dormant. For others, it won’t even matter.

At the same time, what is happening is that we are building new pattern recognition engines that facilitate the fusion of human thought and expression, forming new ecologies of knowledge and creativity. That doesn’t mean they’re necessarily ecologically sound—it will take work to make sure they lead us closer to the sunny outcomes proponents think are possible. The Creative Singularity is an invitation to contributors from all disciplines, not only in science and technology, but also in the humanities and beyond, to participate in shaping an emerging intellectual and creative environment, the new reality.

Source of information: compiled from 8BTC by 0x Information.Copyright belongs to the author, without permission, may not be reproduced

Related Posts