Am I the only artist confused by image AIs? Where do I stand? My two worlds have exploded.
There is no question that image/visual AI have completely changed the game for artists.
I strongly believe that the technology is here to stay. I acknowledge that the ai systems are built off of copyrighted datasets.
As someone who works on both sides of the fence: the artist & the programmer… I am so conflicted with this debate.
i’ve been itching to try this: chatgpt prompt > write a chapter about the issues of image AI tools using copyrighted data sets
The use of copyrighted data sets to train image AI tools has been a contentious issue in recent years. One of the main concerns is the potential for copyright infringement, as the use of copyrighted images without permission can be illegal. Additionally, there are concerns about data privacy and the potential for misuse of personal information contained in the images.
One of the key issues with using copyrighted images to train AI tools is the question of who owns the rights to the images. In many cases, the creators of the images may not have intended for them to be used for AI development, and may not have given permission for their use. This can lead to legal disputes over the ownership and use of the images.
Another concern is the potential for misuse of personal information contained in the images. Many images contain sensitive information about individuals, such as their names, addresses, and financial information. If this information is used to train AI tools, it could potentially be used for nefarious purposes, such as identity theft or financial fraud.
Data privacy is also a major concern when using copyrighted images to train AI tools. If the images are not properly secured, they could be accessed by unauthorized parties, who could then use the information for their own purposes. This could lead to serious privacy breaches and potential legal issues.
In order to mitigate these issues, it is important for companies and developers to obtain permission from the owners of the images before using them to train AI tools. Additionally, appropriate measures should be taken to protect the images from unauthorized access, and to ensure that personal information is not misused.
Another approach to avoid these issues is using synthetic images, which are artificially generated images that are similar to real-world images but do not contain any real personal information. This method can be used to train AI models without the risk of infringing on anyone’s rights or privacy.
In conclusion, using copyrighted images to train AI tools raises a number of legal and ethical concerns, including potential copyright infringement, data privacy, and misuse of personal information. To avoid these issues, it is important for companies and developers to obtain permission from the image owners, and to take appropriate measures to protect the images and ensure that personal information is not misused. Additionally, the use of synthetic images can be an alternative approach to avoid these issues.
My experience in copyright data:
My career started in the print industry back in the day of good ole paste up. Images used in advertisements were sourced from a catalogue, shelves upon shelves of magazines which housed stock photography (and clip art). As the internet gained popularity, these catalogues became digitized and the rules for copyright infringement became blurred. It was very common to source images found anywhere online and alter them (enough) to bypass any copyright. in more recent years, the internet has been flooded by profiteers of products depicting copyright material.
As someone who works with images daily, it is still a common client (& sometimes employer) request… to just grab an image from online and blindly use it. Several times throughout my career I have had to say “no” & explain copyright, which is generally met with contempt for my refusal to comply.
it feels like the conversations about the abuse of copyright images have only just entered mainstream awareness… and now the issue is shoved to the forefront with the advent of image AI’s.
The launch of social media:
As a student of art, I was taught to read those user agreements when posting to 3rd party social media websites. Posting art to instagram & Facebook were big no nos because in their fine print it was stated that by uploading your image, it now belonged to them (therefore your individual rights to the imagery were forfeited)… To this day, i avoid uploading my work to 3rd party services for this reason. User license terms and agreements are constantly changing and I don’t have the patience to read through every iteration. I think many users click through, as I do.
User data – what’s the big deal?
I am not a fan of google. Anyone who knows me… knows I roll my eyes the second I hear the word google. Of course I’m a hypocrite, I use the search engine as my primary search engine tool & YouTube has become my primary social media platform. google is a massive marketing tool. from gmail to chrome to maps & website tracking codes… the pinnacle of targeted advertising. i’m speaking of google simply to illustrate how we unintentionally give away our data to be used at a 3rd party’s discretion. Image AI’s are a good example of what can be possible by manipulating these collected datasets.
A heated debate in the case of image AI’s, is that their datasets are based off of copyrighted material, which is only a piece of a larger picture. steven zapata articulates the systemic issues with image AI’s, I strongly recommend that you take 40 minutes out of your day to watch the following video:
Currently, the conversation is tailored to the artist because AI technologies are hitting our industry first… but don’t forget about the viewer & how easy it is to be fooled by the “fake” & misinformed so this issue is also affecting the content you are consuming right now. AI is here.. and it can be applied to any industry.
herein lies my confusion…
The technology is here. We can fight against it, as much as we can, but it’s here. As much as I don’t agree with how it’s built, I still have to learn how to use this technology as a tool. The programmer/computer buff in me is in awe of the technology… I am so impressed that we have come so far and that I will live in an era where AI technology exists and is not just fantasy. I am a strong believer in letting computers & technology automate work so that as humans we can evolve our efforts. Personally, I fail to see how we can improve AI technology unless we learn & use the technology. I believe it is imperative as an individual to work with AI’s while the technology is still in its infancy… before the conglomerate makes it impossible to instill change.
Here’s the thing, the artist’s battle against image AI’s is only the beginning of a bigger conversation… and I recognize this but I believe that boycotting AI is futile & learning how to use the technology is imperative to implement change.
watching this issue spread like wildfire, has me caught in the middle and I will be keeping a close eye on the debate as it progresses.
my cursory contribution to the conversation would be that first and foremost all of these AI technologies should be open source & free to use by anyone. Charging individuals to use a technology based on datasets that were taken from them without consent… feels very wrong in and of itself.
on the flip side (to be considered for a moment)… anything we post online is “published” information. the AI’s job is to scrub through all of that and return the closest match to answer what we have asked it to do. if we are prompting the image AI, …”in the style of Van Gogh” then we are asking for the AI to scrub for copyright data in return (sampling from the original artist). I think one way to be considerate of copyright is to pay even closer attention to the prompts that we’re inputting.
this is a real dichotomy which affects the industry i work in. it’s a topic that pushes forth ethics, litigation & policy. the discussions we are having now can have a direct affect on what happens when AI is introduced into other industry & technology.
I realize that making such a bold statement does come with consequence by perpetuating circumstances but it also opens the door to more conversation resulting in action.
Anyhow, this is an ongoing discussion, let me know what you think.