The use of AI image generation models has not only gained popularity but raised concerns surrounding potential misuse when it comes to training data, including copyright-protected material.
Text-to-image models have gained significant popularity due to their ability to generate diverse, realistic-looking images from just a short prompt. As these models are trained on vast datasets from various sources, there is growing concern that artists' works, including photographs, paintings and other creative pieces, may be used in training without their consent.
To protect their work from being exploited by emerging technologies, artists have turned to two prominent tools known as Glaze and NightShade.
The tools work by adding subtle, invisible distortions (known as poisoning perturbations) to digital images. These are designed to confuse AI models during training. Glaze takes a passive approach, hindering the AI model's ability to extract key stylistic features. NightShade goes further, actively corrupting the learning process by causing the AI model to associate an artist's style with unrelated concepts.
But a team of international researchers discovered these tools have critical weaknesses that cannot reliably stop AI models from training on artists' work.
Murtuza Jadliwala, UTSA associate professor in computer science and core member of the MATRIX UTSA AI Consortium for Human Well-Being , worked alongside Hanna Foerster from University of Cambridge and Sasha Behrouzi, Phillip Rieger and Ahmad-Reza Sadeghi from the Technical University of Darmstadt to show the insufficiency of existing copyright protection tools and the need for more robust approach.
The team developed LightShed, a powerful new method capable of bypassing these protections. LightShed can detect, reverse-engineer and remove the distortions, effectively stripping away the protections and rendering the images usable again for generative AI model training.
LightShed works through a three-step process. It first identifies whether an image has been altered with known poisoning techniques. In the second step, reverse engineering takes place as it learns the characteristics of the perturbations using publicly available poisoned examples. Finally, it eliminates the "poison" to restore the image to its original, unprotected form.
In experimental evaluations, LightShed successfully detected NightShade-protected images with 99.98% accuracy and effectively removed the embedded protections from those images.
"This shows that even when using tools like NightShade, artists are still at risk of their work being used for training AI models without their consent," Foerster said.
Although LightShed reveals serious vulnerabilities in art protection tools, the researchers stress that it was developed not as an attack on them — but rather an urgent call to action to produce better ones.
"We see this as a chance to co-evolve defenses," Sadeghi explained. "Our goal is to collaborate with other scientists in this field and support the artistic community in developing tools that can withstand advanced adversaries."
The team cites a need for stronger, more adaptive defenses within the rapidly growing landscape of AI and digital creativity.
In March, OpenAI rolled out a ChatGPT image model that could produce artwork in a style reminiscent of Studio Ghibli, the renowned Japanese animation studio. This sparked a wide range of viral memes — and equally wide discussions about image copyright, in which legal analysts noted that Studio Ghibli would be limited in how it could respond to this since copyright law protects specific expression, not a specific artistic "style."
Following these discussions, OpenAI subsequently announced prompt safeguards to block some user requests to generate images in the styles of living artists. But issues over generative AI and copyright are still going on, as highlighted by the copyright and trademark infringement case currently being heard in London's high court. Global photography agency Getty Images is alleging that London-based AI company Stability AI trained its image generation model on the agency's huge archive of copyrighted pictures. Stability AI is fighting Getty's claim and arguing that the case represents an "overt threat" to the generative AI industry.
"The increasing number of lawsuits filed by media corporations against generative AI service providers underscores the seriousness of the unauthorized use of copyrighted content for training AI models. By demonstrating how current protective measures for artists and their copyrighted works can be circumvented, our aim is to establish a strong foundation for the development of more robust and effective defense mechanisms to address this growing concern," Jadliwala said.
The team's study has been accepted for publication at the USENIX Security Symposium 2025 , a premier cybersecurity and privacy conference.
