Whether you hate or love AI art generators, the truth is that, one thing is to feed the algorithm with the works of long-gone artists, and another is to use images done by living artists without their consent.
In order to learn how to create an image in response to a text prompt, AI art generators go through a period of "training." To finally produce an accurate image of real-world objects, colours, and scenes—along with art techniques and style—it has to examine hundreds of millions of image-text pairs as part of the learning process.
As it turns out, AI models must absorb human creativity. For instance, the open-source LAION-5B dataset, which contains billions of photos from all over the internet, was used to train two AI art generators, such as Stability Diffusion, OpenAI's DALL-E and Midjourney.
Although there are talented artists that incorporates AI into their practice in a way that the end result has nothing to do with the images used to train AI (and so they use AI in the same manner as late 19th-century painters employed photography in their works), If anyone is allowed to produce an image "in the style of" an artist, the market may become saturated with imitations, raising questions about authenticity and raising the possibility of financial loss.
If you want to know more about my opinion on the AI art debate, you can watch my video here, and if you're worried about datasets scraping your images to train AI, you can start using the following methods to protect your art. 1. Opt out of training data for AI If you suspect your work has been used without your consent, you can now verify this thanks to new platforms that can tell you in a matter of minutes if your art has been fed to an AI. One of these tools is Have I Been Trained?, a search engine created by an artist collective called Spawning that lets you know if your photo or artwork is one of 5.85 billion original pieces in the LAION-5B dataset used to train Stable Diffusion and Google’s Imagen. They subsequently added the possibility to opt out of the dataset and LAION has stated that it will comply with their request and take the requested images out of its archive. Despite the opt-out tool by Spawning still needing some development, as at the moment it is not possible to include several works at once, this is a huge step forward in allowing artists to have more control over their work shown online. 2. Copyright your art formally
Although any work of art you create is automatically copyrighted as soon as you complete it, if someone has used your artwork without your permission and you want to strike back at court, you will need to prove that the piece in question is actually yours, which will delay the whole process as you’ll need to find for proof that you had actually created it. That is, though copyright may not be up-to-date with AI technology, it can still add to your defence moving forward. In this case, the best thing to do is to formally copyright your artwork and pay the copyright fee, which in the US costs around $35 and in the UK £46. After registering it, you’ll become the de facto author of your work. Check out my blog on why you need to copyright your art for an in-depth look at how it all works.
3. Use Robots.txt to block website crawlers Website crawlers are tools that assist search engines like Google in finding and indexing the most pertinent data to show off on its search results page. That is, they crawl across the Internet searching for specific information. As you may have guessed, some of them are employed to gather images from websites for inclusion in databases for AI training. If you don’t want particular items of your website, like images of your work, to be crawled and subsequently indexed then you should use something known as a robots.txt file. This text file restricts all or most access to certain areas of your website, letting search engines’ web crawlers used to feed AI know what they can and cannot scan and so it effectively deters them from scrapping your artworks. If you want to try this method, ask your web developer to place a Robotos.txt in the backend code of your website to avoid having your images taken and included in an AI training database.
Hope you find this useful and let me know if you want more content related to copyright and how to protect your work!
Commentaires