An artist has come up with a novel anti-theft script that incorporates a watermark identical to that used by AI creators to safeguard their creations.

As the confrontation between humanity and technology escalates, folks find themselves arming up with the same AI tools in a bid to outsmart it.
The news was posted This was shared on Artstation by a user known by their online alias. Eballai .

The artist crafted a Python script that discreetly adds an invisible watermark to PNG images, rendering them 'flavorless' for hordes of AI bots that scavenge the internet seeking new fodder for their algorithms. This announcement reignited excitement within the online community.

How does it work?

AI generators embed an invisible watermark in the images they create using a Python module. Since the watermark is embedded within the image rather than simply overlaid, it goes unnoticed by human observers, yet AI bots can detect it right away during their internet searches for data. This watermark acts as a filter, blocking any AI-generated images from contributing to future AI training.

Experts claim that all image creation software, including popular ones like Midjourney and DALLE2, embed watermarks in their output. AI crawlers, which scour the web for images to enhance their algorithms, avoid these watermarked images, thus preventing them from being incorporated into their training datasets.

The backstory surrounding Ebalai is rich with details. Once the news broke, coders flocked to GitHub to dissect the code. It turned out to consist of just 12 lines of Python, largely a duplication of the code responsible for watermarking generated images. Stable Diffusion It’s amusing that while this is ostensibly a project about artist rights, the developer was quick to copy-paste the code without crediting the original source.

Ebalai, recognized as a formidable opponent of AI, is none other than the artist Edit Ballai from New Zealand. She has since removed all of her works from Artstation. Here’s a link to her discussion around the watermarking issue. Stable Diffusion 2 ChatGPT’s watermarking capabilities could potentially assist Google in identifying text generated by AI.

Microsoft and Google are poised for a major clash in 2023, setting the stage for an intensely competitive year among the giants of AI. link , please keep in mind that the information presented on this page is not meant to serve as, nor should it be seen as, legal, tax, investment, or financial advice. It’s crucial to invest only what you can afford to lose and to seek independent financial counsel when in doubt. For more detailed insights, we recommend consulting the terms and conditions along with the support pages provided by the respective issuer or advertiser. MetaversePost values accurate and unbiased reporting, though it’s important to note that market conditions can change unexpectedly.

Read more about AI:

• Damir oversees the team at Metaverse Post, where he works as a product manager and editor, focusing on emerging topics such as AI/ML, AGI, LLMs, Metaverse, and Web3. His writings draw in an audience exceeding a million readers each month. With a decade of experience in SEO and digital marketing, Damir has been featured in major publications like Mashable, Wired, Cointelegraph, The New Yorker, Inside.com, Entrepreneur, BeInCrypto, among others. As a digital nomad, he travels extensively between the UAE, Turkey, Russia, and other CIS nations. With a bachelor’s degree in physics, he credits his education with providing the analytical tools necessary to thrive in the rapidly evolving online landscape.
• Polygon has unveiled the 'Agglayer Breakout Program' aimed at fostering innovation and delivering value to POL stakers through airdrops.
• Jupiter DAO has introduced its proposal 'Next Two Years: DAO Resolution', which focuses on achieving progressive independence and securing high-level funding.