Matters Journal

View Original

Generative AI Will Change The World: Part 1

They Grow Up Fast. Daniel Simons. Created in Midjourney.

Words by Daniel Simons

In his book, Homo Deus, author and historian Yuval Noah Harari argues that the rise of the ‘useless class’ is one of the most dire threats of the 21st century. The Democratic candidate for the 2020 American elections, Andrew Yang, was so concerned about automation that he ran his campaign on a platform championing Universal Basic Income.

The fear that exponential technological evolution poses an existential threat to the livelihoods of millions is not new, but the tech evangelists of recent past used to mollify their critics by arguing that innovation would take away the ‘bullshit jobs’ and free up humans to focus on the things that matter, like leisure, meaning and creativity.

Now, as the Generative AI explosion engulfs the world - almost without warning - it is clear that the robots are coming for the fun stuff too.

Suddenly everyone is on notice.

Science Fiction writer Arthur C. Clarke said that any sufficiently advanced technology is indistinguishable from magic, and Generative AI certainly feels magical. Even its creators don’t completely understand how it works.

While the technology might be unfathomable, the revolutionary impact it will have is undeniable. Pandora's box is open, but the shape of the future is still being decided. By leveraging the power of AI we could unlock unprecedented levels of human creativity and innovation, regenerate the world, and usher in a new age of abundance.

Or, if we are not careful, Generative AI could lead us to a world of redundant humans and unimaginable inequality. The only certainty is that we have entered a pivotal moment for Artificial Intelligence, and for humanity, and it will continue to evolve at a dizzying pace.

The ‘genie is out of the bottle’ and we all need to adapt or get left behind.

"While the technology might be unfathomable, the revolutionary impact it will have is undeniable. Pandora's box is open, but the shape of the future is still being decided. By leveraging the power of AI we could unlock unprecedented levels of human creativity and innovation, regenerate the world, and usher in a new age of abundance."

What is Generative AI?

Metaverse dreaming. Daniel Simons. Created in Midjourney

The two main technologies driving the Generative AI revolution are image generation programs, and large language models.

Image programs and LLMs are typically trained on enormous data sets. For example, the most popular LLM, GPT3, (Generative Pretrained Transformer 3) was trained on over 1 trillion words, which included billions of web pages, millions of books, and all of Wikipedia.

I asked Open AI’s ChatGPT to explain Generative AI to me. This is the answer it returned within seconds:

*"Generative AI refers to a type of artificial intelligence that is able to generate new content, such as text, images, or music, based on a set of examples or a specified model. This is achieved using techniques such as deep learning and machine learning, which allow the AI to analyse existing data and generate new content that is similar in style or structure."*

Existentialism is a Humanism. Daniel Simons. Created in Midjourney

Generative AI has been with us for years, but recent ease of access and usability has seen it go truly mainstream. A major breakthrough came in June 2020 when Open AI, a company co-founded by Elon Musk and funded by Microsoft, released GPT-3. 

In 2021 the Open AI team then trained a model on billions of pairs of images and text descriptions and unleashed a text-to-image program called DALL-E, which was quickly updated in July 2022 with DALL-E2. 

Since then, there has been an explosion in Generative AI applications and integrations. Outputs from programs like GPT3, Midjourney, Lensa, DALL-E 2 and Stable Diffusion have been flooding social media channels. Canva, Notion, Microsoft, Figma and Photoshop have all announced Generative AI plugins and integrations.

ChatGPT was so popular that it amassed over one million users in under five days and surpassed 100 million active users within two months. AI can create images, music, 3D sculptures, animations, product designs and video games. At the rate AI is sweeping the world it won’t be long before every touchpoint of our lives is enmeshed with some form of hyper-intelligent algorithm. 

Solar powered flying koala. Daniel Simons. Created in Midjourney

The advancements of AI are already being implemented in ways that would have been almost unimaginable until recently. James Earl Jones, the actor who played Darth Vader, has sold the rights to his voice to Ukrainian startup Respeecher.

The film ‘Fall’ was given an R rating because it contained too many F- bombs, but the producers didn’t have the budget to reshoot. The solution: they used the DeepFake company Flawless to alter the actors’ performances to seamlessly replace the F-bombs with PG-friendly expletives. Soon the same technology will be used to seamlessly dub any film into foreign languages.

Thanks to Generative AI, you can now listen to a podcast conversion between AI bots of Steve Jobs and Joe Rogan, or you can tune in to a completely fabricated ‘infinite conversation’ between Werner Herzog and Slavoj Žižek.

AI is also being used for coding, drug discovery, policy development and almost anything that can benefit from processing large amounts of data.

If the lightning pace of these advancements is causing your mind to swell with anxiety, fear not, you can always book a session with the AI psychologist Woebot, or talk through your problems with one of Replika’s AI companions.

"AI is also being used for coding, drug discovery, policy development and almost anything that can benefit from processing large amounts of data.

If the lightning pace of these advancements is causing your mind to swell with anxiety, fear not, you can always book a session with the AI psychologist Woebot, or talk through your problems with one of Replika’s AI companions."

What is Generative Art?

Portal To The New Now. Daniel Simons. Created in Midjourney.

Generative AI art can refer to any art that is created using artificial intelligence algorithms. Paintings, sculptures, music and even literature can already be created with AI.

There has been a deluge of applications that can create mind-blowing Generative Art. The most popular are DALL-E 2, Midjourney and Stable Diffusion. Google has a program called ‘Imagen’ in development and Meta has created its own text-to-video application.

The programs have become viral because they are so enchanting and so easy to use. Often all you need to do is upload an image or type in a text prompt and then, within seconds, the picture-bots will produce an uncannily realistic or wonderfully aesthetic work of art. Ilya Sutskever, a co-founder at Open AI, describes the process as ‘transcendent beauty as service.’

The City is Regenerative. Daniel Simons. Created in Midjourney

The World-Changing Potential

While social media feeds are blowing up with the self-gratifying output of experimenters and tinkerers, the new tools are already sending shockwaves through the creative and communications professions. Programs like Jasper AI, Writesonic, Rytr, Copy.ai and ChatGPT are being used to create everything from blog posts to news articles. 

Tools like Runway and InVideo are enabling text-to-video editing, Aiva and Soundraw are  opening the floodgates for synthetic music. You can even create your own video avatars with Synthesia.io or Synths Video. Further down the track, AI will be used to create films and videogames, or build virtual worlds in the Metaverse. Pretty soon we could all be living in a ‘text-to-anything’ era limited only by our imaginations. 

Regenerative utopian future city. Daniel Simons. Created in Midjourney


Artists, graphic and industrial designers, writers, and marketers are all starting to grapple with the implications. Generative AI is also fundamentally changing the way some architects develop their projects. Conceptual designs that could have taken days to draw, model, and render can now be generated in minutes. For example artists like Tim Fu are creating stunning prototypes by using combinations of Midjourney, DALL-E 2 and Photoshop. 

According to Venture Capital firm Sequoia, the current limitations of Generative AI mean that professionals are mostly using it for prototypes or first drafts, but given how new this technology is and how quickly it is improving, it won’t be long until the AI programs are capable of creating final products. 

The Controversies and Conundrums

No Respite. Daniel Simons.

The backlash to these new technologies was inevitable. When Jason Allen won a Colorado Art Fair for his "Théâtre D'opéra Spatial," which he created with Midjourney, the Twittersphere blew up, lamenting ‘the death of art.’ Artists' groups and platforms are quickly banning AI art. Almost every artistic peak body or union is preparing for battle. Even actors’ unions have been thrown into the fray because AI is stealing celebrities' faces and voices to create synthetic content almost indistinguishable from reality. 

AI is already being used in the mainstream media. Cosmopolitan was the first magazine to use AI art on its cover. It took 20 seconds to create. In August 2022 an article about Alex Jones in The Atlantic used an image credited to Midjourney. 

Readers were furious that such a well-resourced publication opted for an AI-generated image rather than pay a human artist. The outrage was so fierce that it forced the journalist, Charlie Warzel, to issue a public apology. The Guardian commented by saying that the outcry served as ‘a salutary warning that publishers who give work to machines rather than creative artists deserve everything they get’. 

It’s not just the use of the tools causing concern, the technologies themselves have also been called into question. Generative AI has been severely criticised for being biassed, reinforcing racial and gender stereotypes, for outpacing copyright laws, and for the wide array of nefarious uses it enables. 

Parametric water fashion shoot in the future. Daniel Simons. Created in MidJourney.

Bias is a problem with almost all Generative AI programs and the issue is acknowledged by most of the program creators. In an interview with WIRED magazine, the DALL-E team acknowledged that if you asked their program to draw a CEO, you would likely get an image of a man. If you asked it to create an image of a nurse, you’d get a woman.

Dark skinned individuals were most likely to be associated with words like ‘prison’ or ‘angry' and negative images were associated with some religions. Women who uploaded their selfies to the Lensa AI app for artificial stylisation were horrified to find it ‘undress or sexualise’ them without their consent.  

More concerning and more difficult to remedy than bias, is the copyright quagmire. The relationship between Generative AI and ownership is a hot mess. Art created by AI often can’t be adequately protected by copyright. 

Artists who use Generative AI programs to create their work are at risk of unwitting copyright infringement, and the artist’s who’s styles and creations are used to feed the algorithms are not being compensated. Often they have no way of opting out of being exploited.

Generative AI datasets are derived from an enormous pool of images, a vast quantity of which are copyrighted. Many of the software companies argue that their models rely on ‘fair use’ but the copyright laws have been described as ‘murky’ at best. 

One of the biggest concerns is that current copyright laws only protect artists from use of their actual work. They don't protect them from works inspired by their style. The lawmakers didn't anticipate a world where anyone with a basic grasp of English could create thousands of perfect-quality derivative works simply by typing 'in the style of …' into a prompt box. This is an existential threat to artists that current copyright laws are not equipped to deal with. 

Greg Rutkowski, an artist who is renowned for his distinctive style of dragon-inspired art, has claimed that his name has been used to create over 93,000 images using Stable Diffusion. He used to have a minimal internet presence. 

Now if you search for his name you will find thousands of images in his style that were not created by him. He has no control over it and receives no compensation.  

Regulations are bound to change, but nobody knows what that will look like. Getty Images has begun legal proceedings against Stable Diffusion and a class action lawsuit has already been filed in California against Stability AI, Midjourney and Deviant Art by a collective of artists. 

The suit was filed by the Joseph Saveri Law film, who are also currently suing Microsoft, Github and OpenAI in relation to Copilot, the automatic code generator. The outcome of these lawsuits could have deep ramifications for the future of Generative AI.

Two women playing tennis on a turtle made of solar panels. Daniel Simons. Created in Midjourney

Another vital concern is the concentration of power and the radical redistribution of income from the many to the few. Multi-billion dollar companies like Open AI, Google, Microsoft and Meta are hoovering up all of the creative outputs of humanity and repackaging them for a windfall in subscription fees. 

The wealth trickles up to the top and the creators get left in the dirt.  The process has been described by The Next Web writer Tristan Greene as akin to ‘Soylent Green,’ in reference to the Sci-Fi film where humans are unwittingly forced to devour other humans for sustenance. 

As these new technologies advance they also put god-like powers into the hands of anyone who wants to manipulate reality for nefarious ends including identity theft, scams, deep fakes, and fake news. Google's Generative AI program, Imagen, has been withheld from the public due to concerns over the potential risks of misuse. 

Google was worried that its datasets included too much pornography, social stereotypes, and other harmful or derogatory content. The DALL-E 2 team was also concerned with deep fakes, they prohibited users from asking the AI to generate images that depicted famous people, violence, or political content. Other programs like Stable Diffusion are open source, meaning anyone can be altered or manipulated by anyone. 

As enormous as the disruption caused by Generative Art may be, it will be utterly dwarfed by the revolutionary impact of LLMs. ChatGPT has passed Law exams, received a B grade for an essay at the Wharton Business School, and is being used by school and university students to write essays. 

It has already been used to write news articles, entire non-fiction books, and novels. The impacts on the education system and any profession that involves thought or the written word are difficult to imagine. Especially as new programs like Google’s ‘Bard,’ and other upgrades based on GPT-4 are released into the wild.

The Future

The Golden Age of Future Now. Daniel Simons.

The era of Generative AI is here and along with the magical picture-bots, synthetic media, and endless streams of bizarre cat photos, it has brought with it a dizzying sense of wonder and a deluge of questions.

What impact will it have on the global workforce? Will artists and creatives be able to survive without embracing the new tech? How will we solve the challenges of copyright, ownership, and equality? What is the future of art? What is the future of education? 

As technology evolves faster than our ability to integrate it into our understanding of the world, we will be forced to answer even more perplexing questions: How will we find meaning when reality can be so easily manipulated or fabricated? Will an endless torrent of ‘on-demand’ art lead to beauty fatigue and ennui? 

Will we all end up living off our universal basic incomes, plugged into AI generated metaverses, communing with artificial beings?  Will we reach a point where every possible permutation of art and idea has been rendered and left us with no space to dream?

In the future, we might all be forced to join the ranks of the ‘useless class’, but for now, we'll have to concern ourselves with the questions of using and being used. How can we use AI to create the world we want, and how do we want to let it use us? 

At the pace things are advancing, we'll need to move quickly.