Artists are educating fans and the public, and resisting unauthorized use of their copyright protected artWORK.
Artist Karla Ortiz started in blog in late 2022. It documents her journey on the vanguard of raising awareness about AI art impact on artists: https://www.kortizblog.com/
The Concept Art Association Go Fund Me "Protecting Artists from AI Technologies" has worldwide support. You can read more about it here: https://www.gofundme.com/f/protecting-artists-from-ai-technologies?utm_campaign=
and the Concept Art Association website: https://www.conceptartassociation.com/about
Some older blog posts about why artWORK is not free:
https://stuartngbooks.blogspot.com/2016/01/art-has-value-raising-awareness-about.html
https://stuartngbooks.blogspot.com/2017/08/inspiration-imitation-art-theft.html
---------------------------------------
UPDATE 3.20.23
University of Chicago professor Ben Zhao featured in this 3.17.23 link from TechCrunch https://techcrunch.com/2023/03/17/glaze-generative-ai-art-style-mimicry-protection/?fbclid=I
“What we do is we try to understand how the AI model perceives its own version of what artistic style is. And then we basically work in that dimension — to distort what the model sees as a particular style. So it’s not so much that there’s a hidden message or blocking of anything… It is, basically, learning how to speak the language of the machine learning model, and using its own language — distorting what it sees of the art images in such a way that it actually has a minimal impact on how humans see. And it turns out because these two worlds are so different, we can actually achieve both significant distortion in the machine learning perspective, with minimal distortion in the visual perspective that we have as humans,” (Zhao) tells us.
“This comes from a fundamental gap between how AI perceives the world and how we perceive the world. This fundamental gap has been known for ages. It is not something that is new. It is not something that can be easily removed or avoided. It’s the reason that we have a task called ‘adversarial examples’ against machine learning. And people have been trying to fix that — defend against these things — for close to 10 years now, with very limited success,” he adds. “This gap between how we see the world and how AI model sees the world, using mathematical representation, seems to be fundamental and unavoidable… What we’re actually doing — in pure technical terms — is an attack, not a defence. But we’re using it as a defence.”
Another salient consideration here is the asymmetry of power between individual human creators (artists, in this case), who are often producing art to make a living, and the commercial actors behind generative AI models — entities which have pulled in vast sums of venture capital and other investment (as well as sucking up massive amounts of other people’s data) with the aim of building machines to automate (read: replace) human creativity. And, in the case of generative AI art, the technology stands accused of threatening artists’ livelihoods by automating the mimicry of artistic style."
--------------------------------------
Karla Ortiz debuted her "Musa Victoriosa" art -- and imporant news about Glaze -- on her Facebook page on March 15th:
"Today
I am happy to announce what could be the most important oil painting I’ve made:
Musa
Victoriosa
This
is the first painting released to the world that utilizes Glaze, a protective
tech against unethical AI/ML models, developed by the The University of Chicago
team led by Professor Ben Zhao.
The
app is out NOW and you can download it here:
https://glaze.cs.uchicago.edu/download.html
This
painting is a love letter to the efforts of this incredible research team and
to the amazing artist community. This transformational tech takes the first of
many steps, to help us reclaim our agency on the web, by making our work not be
so easily exploited.
So
how does Glaze work?
Professor
Zhao describes:
“Glaze analyzes your art, and generates a modified version (with barely visible changes). This "cloaked" image disrupts AI mimicry process.” (Source of tweet:https://twitter.com/ravenben/status/1636131323847868418...)
For
a more in-depth description visit here: https://glaze.cs.uchicago.edu/index.html#whatis
Best
part about Glaze? It’s free for everyone and the researchers will be working
hard to bring even more updates to this badass tool! So my painting can look
even closer to the source amongst other features. So share far and wide!
Personally
I’ll be shutting down parts of my personal art website for a while, while I add
Glaze to the works featured. I’ll also
be integrating Glaze as a final touch for all future works posted online.
The
tech is already good now and it’s only going to get better! 🙂
Once
again special thanks to the incredible team from The University of Chicago:
Shawn Shan, Jenna Cryan, Emily Wenger, Heather Zheng, Rana Hanocka and
Professor Ben Zhao.
Your
work gives me all the hope! I truly believe your efforts will become the beacon
of many and truly change our wild world for the better!
A
HUGE thank you to Lynds Gallant for making the ui/ux not only look great but
easy to use!
Thank
you to Nathan Fowkes Art for also helping develop this and to all artists who
took the survey!
Thanks
to all the press including Kashmir Hill for covering this incredible tool too!
And lastly a special shout out to my brushes, my oils, my wacom and my old version of Photoshop (did final touches there). Actual artists tools that don’t rely on ill gotten data and exploitation to work just fine. 😉
----------------------------------------
Here's more info on Glaze from the links shared in Karla's post above.
Download
Glaze (Beta) https://glaze.cs.uchicago.edu/download.html
"First,
thank you for your patience. We've been busy working on packaging Glaze into
the most intuitive and usable app for artists. Please note that by downloading
Glaze, you are acknowledging and agreeing with our software license.
After
the initial download, the app will download additional ML libraries and
resources that will require stable Internet access and approx. 4GB of storage.
Before
using Glaze, be sure to read the User's guide and FAQ, which includes step by
step instructions on glazing your art and answers to common questions you might
have.
Please
note that this is clearly an early release. Please be patient if you run into
issues. While we are not software engineers, we will do our best to improve
Glaze with your feedback. You can submit both comments and suggestions as well
as bug reports.
OS-specific
notes.
For
Windows users, we are in the process of "registering" the Glaze app
with Windows. Getting these signature keys is taking longer than we expected,
so Glaze.exe is currently signed with a non-EV key. And you will likely get a
warning when you install Glaze for the first time. This should go away as more
users download Glaze and the our developer key gets a higher reputation as a
result with Microsoft.
For
Mac Intel users, the app is currently limited to machines running MacOS 13 and
higher. We will look into supporting older MacOS versions, but did not want it
to delay the Glaze release."
and this...
Glaze
Software User Guide https://glaze.cs.uchicago.edu/index.html#whatis
"Welcome to Glaze! This user guide will walk you through the steps to protect your artwork with Glaze. This is a living document, and we will continuously add content to the guide as we hear back from artists and their experiences.
First, if you haven’t already, please read more about Glaze and its risks and limitations before using the tool on your artwork. Second, this is an early release. Please be patient as we fix issues and improve the tool over time. If you run into issues, you can file a bug report.
Step
by Step Guide
The
first time you install Glaze, it will need to download a number of machine
learning libraries and pretrained models. This might take some time,
particularly if you are outside the USA. Note that these are one time
downloads, and do not need to be fetched again if you update the Glaze app.
Once Glaze is fully installed, follow these instructions below:
Please run Glaze as the last step before you post your artwork online. If you do any image resizing, format conversion, or watermarking, please do so before you run Glaze. This will maximize the protection effect.
Step 1: Select your artwork. You can input a single image by dragging it into the image placeholder or selecting multiple images using the "Select..." button.
Step 2: Select Glaze parameters.
Intensity: Specify the intensity of "the style cloak" applied to your artwork. Higher intensity leads to stronger protection, but also generally correlates with more visible changes to your artwork. The more Glaze changes your artwork with perturbations, the harder it will be for AI models to recognize your original artistic style, making it that much harder to mimic. We recommend the "Low" level intensity or higher. You can also be creative with the intensity level. For example, you can use different intensities for different distribution channels or higher intensities for less important (e.g. older) artwork.
Render quality: This represents the compute time spent searching for the optimal style cloak given your intensity level. Longer render quality leads to better performance but longer compute time. The highest render quality takes around 60 mins per image on a personal laptop. This will highly depend on the specs of your laptop. GPUs are not required but if present, they will greatly reduce the processing time
Step 3: Preview or Run. Once you upload the image(s) and select your desired settings, you can "Preview" the Glazed result. Preview allows you to see the estimated final image(s) with the Glaze changes, so you can make adjustments to the Intensity level if needed, before rendering the final image(s). Note that the preview result might not be representative of the final result. Once you are satisfied with the Intensity level, you can click Run, which will Glaze the image(s). The resulting image(s) will be saved in your "output directory" with the same file name.
Note: Glaze protection varies with different styles of art. Some artistic styles (e.g., character design, animated art), are more vulnerable to protection removal due to the “smoother surfaces” of the style. As a result, we encourage you to use an Intensity level of "Low" or higher. After the image is modified, Glaze checks the effectiveness of the result, and warns you if it fails to offer enough protection.
Issues and feedback: Please fill out our feedback form to let us know about your experience using Glaze. Please fill out our bug report with any issues.
FAQs
Glaze said it encountered errors on my images, how should I fix it? You can check the details of the errors in error.txt in your output folder. The most common error is Glaze fails to offer strong enough protection to some of your artwork. When this happens, you can try to re-run Glaze on those artwork. If the error persists, you should consider increasing the Intensity level and/or render quality before re-run Glaze on those artwork. As we discussed above, some type of artwork will be harder to protect than others.
Why does the same intensity level look different across art pieces? The same intensity level (pixel change magnitude) may have different noticeability on different types of artwork. For example, the change is more noticeable on smooth or darker surfaces.
How do I know the protection is sufficient/effective? It is challenging to estimate the success of the protection because it would require someone to build a customized AI model mimicking your artwork. We currently do not plan to release such an evaluation tool given that it can easily be misused for malicious purposes.
How much computer resources is Glaze using? Glaze uses around 5 Gb of memory and a substantial amount of CPU computing. We are working on reducing these numbers.
What should I do if Glaze warns insufficient computer memory? If you computer has less than 9Gb of memory, you can still run Glaze but it may take a much longer time to render. You can speed it up by closing any other applications that use a lot of memory. If your computer has less than 5Gb of memory, we do not recommend you to run Glaze currently. We are working on reducing Glaze's memory usage.
Does Glaze generate the same result each time given the same original art? No, Glaze will generate slightly different results given the same original art and settings.
Can I glaze my art twice for better security? No, it could actually hurt the protection. Increasing the Intensity level and/or render quality is currently the only way to improve protection strength.
Will Glaze support other operating systems (Linux, iOS, Android)? Currently, we do not plan to support platforms other than Windows and MacOS. You are welcome to install Windows VM on Linux to run Glaze.
Disclosure:
you will be downloading some of the resources from HuggingFace (a common
platform to host resources) and we use checksums to make sure the files are
authentic and have not been tampered with. However, if you would like to
download from the University of Chicago server (on a slower connection), you
can email us for instructions."
No comments:
Post a Comment