Top 5 #AiArt stories of 2022 and what to watch for in 2023
The year 2022 may stand as an inflection point in human history because of the transformative nature of artificial intelligence and the speed with which it has become easily accessible
In a few years when we look back on 2022, there’s a good chance we’ll remember it as the year everything changed. The dizzying advances in artificial intelligence, its wide-spread accessibility and its impact on all corners of our society feels transformative.
The train has left the station and it has no brakes. Where it takes us from here is equal parts exciting, mind-boggling and terrifying.
Here are my picks for the five most significant stories related to Ai-generated art in 2022. Feel free to let me know in the comments what I missed.
Stable Diffusion, DALL-E 2 and Midjourney started a revolution
How stock image services are handling the advent of AI-generated images
Stable Diffusion, DALL-E 2 and Midjourney started a revolution
What happened: Let this sink in for a second: At the start of 2022, neither Stable Diffusion, DALL-E 2 or Midjourney were known outside of their respective research labs. Today, just nine months after DALL-E 2 was the first of the three unveiled, the platforms are attracting millions of daily users, have been the subject of thousands of news stories, have generated more than a billion images and sparked a furious anti-Ai backlash.
The speed with which generative art is evolving is hard to grasp. From MIT Technology Review:
“The shock and awe of this technology is amazing—and it’s fun, it’s what new technology should be,” says Mike Cook, an AI researcher at King’s College London who studies computational creativity. “But it’s moved so fast that your initial impressions are being updated before you even get used to the idea. I think we’re going to spend a while digesting it as a society.”
Background: While the big three AI-art platforms all hit the scene in 2022, they weren’t the first. ArtBreeder has been around since 2018 and even inspired a prescient 2020 article entitled, Will this creepy AI platform put artists out of a job? You can be forgiven for not noticing because like most of us, you were probably pretty pre-occupied with trying to stay alive.
What changed from 2018 to 2022? The game changer is something called a “diffusion model,” which involves two neural networks that work together to interpret text prompts and generate images. I won’t go into all the details of how diffusion models work but if you’re interested, you can find a comprehensive explanation here.
The impact: By putting such powerful technology at the fingertips of virtually anyone with a smart phone or computer, the AI labs have allowed anyone, regardless of their “artistic” ability, to creatively express themselves in ways they could never have done before. AI has democratized art.
Others look upon it a less benignly. They fear the jobs lost in the design field and creative sectors. They see the disinformation that can be spread to cause harm and suffering. They stand with the human artists who never gave permission – nor received compensation – to have their work used to train AI art-generators.
What to watch: Video. Text-to-video generators will be the next rabbit hole to swallow us whole. Text-to-video is already here but the platforms aren’t ready for prime time. We can expect that to change rapidly. Other than that, how far will the big three and others push the envelope? What kind of guardrails will they add to respond to the complaints from artists and others concerned about the ethics of AI?
Human artists aren’t going down without a fight
What happened: A protest against AI-generated images on the ArtStation platform inspired a movement that has put a spotlight on the debate of how AI-generated images are used and created.
Background: ArtStation, which is one of the leading platforms for artists to share their work, angered users when it featured generative art on the site’s Explore page. Artists protested by spamming their accounts with what has become the symbol of the movement, the letters “AI” inside a red circle with a slash through it. The outcry led ArtStation to clarify its stance on AI-generated images: it will not prohibit AI-generated images but it also warns that AI companies using ArtStation to train their AI may be infringing on the rights of its users.
Why it matters: Coming at the same time that the Lensa selfie phenomena was taking the internet by storm, the artist revolt helped push the debate over the ethics of AI training models and whether they’re ripping off artists without compensation to the forefront.
What to watch: While the anti-AI protest movement won’t derail the growth of AI-art platforms, it may lead to more self-regulation among AI companies, and could lead to some actual legislation or new case law. Stable Diffusion had already moved in that direction with its November update and has announced it will let artists opt out of the training set in its next release. Another development to watch is a planned amendment in the UK that would criminalize distributing non-consensual deep fake porn.
Will ChatGPT take generative art to the next level?
What happened: In November, OpenAI, the ground-breaking research firm behind DALL-E 2, released the free preview of ChatGPT, an incredibly sophisticated chat bot that can research queries and respond in ways that are eerily human-like. It makes well-reasoned and supported arguments, creates elaborate stories from simple prompts, writes poetry in any style, and can even generate computer code.
It also sometimes gets basic facts wrong – often enough that you should take everything it writes with a grain of salt. Yet, play with it for just a short while and it’s easy to see that it has massive potential to be a technological disruptor.
Why it matters: While ChatGPT is not an AI-image generator, it takes a pretty small leap to imagine how OpenAI will find a way to use ChatGPT’s ability to understand language to supercharge the next version of DALL-E. You can actually already ask ChatGPT to write an image prompt for you, you just have to paste the results into your AI-art generator.
Why not have a future version of DALL-E with a chat function that helps you refine your prompt?
What to watch: Many believe that ChatGPT poses a serious threat to Google’s search engine. It’s not just outside analysts but executives inside Google, as well.
When one of the world’s most powerful companies is facing an existential crisis, you can expect it to respond with urgency. Will we see Google’s response in 2023? Of course, the dustbin of history is filled with many companies – MySpace, Blackberry, AOL – that didn’t adapt to the evolving marketplace.
Whatever Google’s next move is, OpenAI is expected to release an updated version of ChatGPT, GPT-4, sometime in 2023.
Crypto Winter brought NFTs down to earth
What happened: In 2021, Mike Winklemann, the digital artist better known as Beeple, sold one of his NFTs at a Christie’s auction for $69 million, making him one of the world’s most valuable living artists. If you somehow didn’t know what an NFT was, that was probably the moment you started paying attention. The headwinds of the Beeple sale pushed the price of collectible avatars and other digital works into the stratosphere and suddenly everyone was a digital artist.
Flip forward today and you’ll find a completely different marketplace. If 2021 was the year that NFT dreams were born, 2022 is where they went to die. Welcome to Crypto Winter.
In early 2022, the value of cryptocurrencies went into a free-fall, with Bitcoin dropping from its November 2021 high of $69,000 per coin to less than $17,000 where it is as I write this. Any hope of a near-term recovery was dashed with November’s FTX scandal and collapse. Not unexpectedly, NFTs have seen a similar crash. OpenSea, one of the largest NFT marketplaces, saw its trading volume drop by 90 percent in 2022.
Crypto and NFTs were struggling well before the FTX scandal. Remember the tweet that Twitter founder Jack Dorsey sold in 2021 for $2.9 million? It was pulled from auction in April when the bidding topped out at $14,000.
One of the most high-profile line of NFTs, the Bored Ape Yacht Club, saw its floor – the lowest price to purchase one – drop from $429,000 in April to about $84,000 today.
Why it matters: The rise of NFTs represented a significant new revenue source for artists. The crypto crash may have put the brakes on much of the wild speculative buying and selling of NFTs but it hasn’t really slowed the creation of unique digital art. The big difference may be that investors can now enter at a much lower price point. And despite the chill, collectors have more choices than ever.
It seems that literally every week, some company announces a new line of branded NFTs. In just the month that I’ve been publishing this newsletter, I’ve linked to former President Trump’s line of collectible trading cards, Starbucks’ new NFT-based loyalty program, the Association of Tennis Professionals’ launch of their LOVE NFT art collection, Timex’s NFT offering to owners of Bored Ape Yacht Club and Porsche’s upcoming NFT project.
What to watch: Expect to see even more NFT releases in 2023 linked to brands mostly targeting gamers, metaverse denizens and digital collectors. Don’t expect a return of the speculative trading that led to the Beeple bonanza. However, the trend of museums and art galleries hosting digital art in person and in virtual spaces should continue and the ownership of digital art will continue to become a mainstream concept.
We can only hope that the Crypto Winter will reduce the number of NFT hucksters on sites like Instagram but that’s probably asking too much.
How stock image services are handling the advent of AI-generated images
What happened: The major stock image providers have taken different approaches to AI-generated art. Getty Images banned AI-generated content, citing concerns about the legality of the images. Adobe Stock went the other direction, allowing AI-artists to sell images with some limitations. Meanwhile Shutterstock is fully embracing AI-art, announcing that it had teamed up with OpenAI, the research firm behind DALL-E, to sell generative images.
Why it matters: Stock image providers have long been an important resource for content creators. News outlets, corporate and nonprofit websites and bloggers often rely on photos and graphics from stock image providers to fill their illustration needs.
How these stock image providers handle AI-generated images will have a big influence on the integration of generative art into the mainstream and will provide a model for how other companies approach AI-art.
What to watch: How successful will Adobe Stock, Shutterstock and others be in selling AI-generated art when almost anyone who can type a prompt can come up with really good illustrations? And perhaps, the question isn’t whether they can sell AI-art but whether their business model can survive at all?
All images were generated by Joe Newman using the Midjourney platform. You can also follow along on Instagram at @most_sublime_media. Get fresh content delivered to your inbox by subscribing to our Substack newsletter.