{"id":494,"date":"2023-05-10T20:50:34","date_gmt":"2023-05-11T01:50:34","guid":{"rendered":"https:\/\/aiforfolks.com\/?p=494"},"modified":"2023-05-10T20:50:39","modified_gmt":"2023-05-11T01:50:39","slug":"does-craiyon-steal-art","status":"publish","type":"post","link":"https:\/\/aiforfolks.com\/does-craiyon-steal-art\/","title":{"rendered":"Does Craiyon Steal Art? What About The Other AI Image Generators?"},"content":{"rendered":"\n
These days, you\u2019d be hard-pressed to tell the difference between some of the art now proliferating the internet that has been created by AI image generators and the real thing.\u00a0<\/p>\n\n\n\n
With new AI art creation engines, such as Craiyon<\/a>, Midjourney<\/a>, and Stable Diffusion<\/a>, being used by multiple millions of people each day, real artists and other creators like photographers all over the world are worried about their copyrighted material being infringed upon.<\/p>\n\n\n\n This is because these artists believe that they have been made unwilling participants in the creation of these AI services, which used their copyrighted material to \u2018learn\u2019 by scraping their images off the internet. <\/p>\n\n\n\n So, does Craiyon steal art<\/strong>? Do the other systems? Well, the answer isn\u2019t as simple as you might think. Read on for an explainer\u2026<\/p>\n\n\n\n Some of you may be old enough to remember the early days of digital music, before the likes of Apple Music and Spotify launched their popular streaming services. All the way back in 1999, along came something called Napster<\/a>, which was a peer-to-peer file-sharing service that people used to share their digital music files.<\/p>\n\n\n\n It was the dawn of the mp3 file, and people uploaded their CDs, as well as rare analog vinyl and cassette tapes, into the format, which allowed anyone with an account to share them. At one point, Napster had more than 80 million users registered, allowing billions of music files to be shared. However, almost all of the music on the site was uploaded without copyright approval.<\/p>\n\n\n\n Napster allowed people to upload their CDs, as well as rare analog vinyl and cassette tapes, into the format, which allowed anyone with an account to share them. At one point, Napster had more than 80 million users registered, allowing billions of music files to be shared. However, almost all of the music on the site was uploaded without copyright approval.<\/em><\/p>\n<\/blockquote>\n\n\n\n So, the Recording Industry Association of America stepped in and filed a lawsuit, and Napster was soon shut down. <\/p>\n\n\n\n This is the same scenario that many artists and photographers are now hoping for, after a class action lawsuit was filed in California, alleging that new AI image-creating software is breaking copyright law by stealing their licensed material.<\/p>\n\n\n\n So, does Craiyon, Midjourney, and Stable Diffusion steal art? Well, if you ask most artists, they believe this to be true. <\/p>\n\n\n\n There has also been a lawsuit filed by Getty Images in the UK, as well as in Delaware<\/a>, against Stable Diffusion\u2019s parent company, StabilityAI, claiming that it misused over 12 million of its photographs to train the AI, and didn\u2019t have a license to do so.<\/p>\n\n\n\n All of these art creation AI models are based on generative artificial intelligences, which are systems that are meant to emulate the technique of the great old masters. The most popular GAI at the moment is DALL-E 2, which is an OpenAI system, now launched as Craiyon<\/a>.<\/p>\n\n\n\n It\u2019s the same company that launched ChatGPT last year, and was co-founded by rocket boy himself, Elon Musk.<\/p>\n\n\n\n And though it seems like these companies have got away with \u2018scraping\u2019 artwork (as they call it) from the websites of the original creators, they do have their own rules when it comes to how users utilize their services. <\/p>\n\n\n\n For example, the prompts that you give these AI art generators are generally successful, to a point. That\u2019s because the developers have implemented controls, such as misuse rules, which won\u2019t allow people to intentionally create images that people would find offensive or distressing, (which is subjective, to say the least).<\/p>\n\n\n\n So, these AI companies have their own rules while making profits on the generation of AI images, even though there are no such rules for the art and photographs that they scrape, which belongs to those they\u2019ve sourced from.<\/p>\n\n\n\n Such is the way of AI, it seems. But let\u2019s take a closer look at Craiyon, Midjourney, and Stable Diffusion to see what they can produce\u2026<\/p>\n\n\n\n Once known as Dall-E Mini, because it is an open-source model based on OpenAI\u2019s creation Dall-E 2, Craiyon is an AI art generator that was developed by Frenchman Boris Day, who produced a more accessible version of the AI art generator for the general public to use.<\/p>\n\n\n\n Probably the easiest prompt you could give Craiyon to start, is to ask it to create an image that is in the style of a particular artist, like Andy Warhol. <\/p>\n\n\n\n So, we tasked Craiyon to produce an image of Taylor Swift by Andy Warhol, and this is what happened:<\/p>\n\n\n\n
\n\n\n\n\n\n\n\nDoes Craiyon Steal Art? What About Midjourney and Stable Diffusion?<\/h2>\n\n\n\n
\n
\n\n\n\nHow Good Is Craiyon?<\/h2>\n\n\n\n