Show summary Hide summary
Child pornography files detected in data collection
An investigation by Stanford Internet Observatory experts revealed the presence of thousands of child pornography files in a widely distributed data collection called LAION-5B. This discovery is particularly worrying given the growing proliferation of child pornography content generated by artificial intelligence on the Internet. The recent analysis carried out by the institution identified files representing sexual violence against minors in data archives intended for training visual synthesis systems. LAION-5B, a repository containing more than 5 billion visuals and explanatory texts from community platforms and adult sites, was used to educate the Stable Diffusion AI developed by the Stability AI company.
Given the ability of AI models to create compelling visuals from just a few examples, the presence of such images within LAION-5B raises concerns about their potential impact. The study authors immediately reported the illegal files to the appropriate authorities, including the National Center for Missing & Exploited Children in the United States and the Canadian Center for Child Protection. They have also taken steps to remove these incriminating images.
Actions taken to contain and prevent these problems
In response to these alarming findings, LAION officials promised to implement a strict policy against the presence of illicit content in their data collection. As a first step, access to LAION was temporarily disabled on the Internet in order to conduct a thorough check and cleanup.
Meanwhile, Stability AI claimed that their AI was trained with a filtered version of LAION-5B. Filtering systems have been integrated to block problematic results during user interactions with AI models. Additionally, identification tools have been put in place to mark AI-generated visuals and prevent malicious exploitation of their technology.
Online child pornography: battle for influence around a controversial European text.
Opponents of the text fear that, beyond the fight for the protection of children, it will open the way to widespread surveillance of private messaging. pic.twitter.com/1dkN4o2nzB
— Jonathan Chan 💡📣 (@ChanPerco) September 26, 2023
Gabriel Fouquet: a writer passionate about technology and AI
Gabriel Fouquet is a writer passionate about technology and artificial intelligence, and more particularly about the way these two fields interact and complement each other. He shares his expertise and discoveries on the blog intercoaching.fr, a site dedicated to technology, AI and their impact on our daily lives.
From a young age, Gabriel was attracted to technology and its influence on the world. His fascination with these fields has grown over the years, from a simple curiosity to a true passion. What has always intrigued him is how these ever-changing fields shape our society and our future.
To read The inevitable technological revolution in robotics: find out why!
By devoting his life to the study and understanding of artificial intelligence and technology, Gabriel aims to demystify these often complex subjects. On the blog intercoaching.fr, he writes articles that translate these concepts into understandable terms, with the aim of making AI and technology accessible to everyone.
His work is driven by a true passion for learning and sharing knowledge. He is constantly on the lookout for the latest advances in these fields and hopes to inspire others to share his curiosity and enthusiasm for technology and AI.