Image Generating AI Models are Being Trained on Explicit Children Images, claims Stanford
In Brief
Popular text-to-image generative AI tools including Stable Diffusion, are being trained on thousands of child sexual abuse images.
Popular text-to-image generative AI tools, including Stable Diffusion from Stability AI are being trained on thousands of child sexual abuse images, as per the research report from the Stanford Internet Observatory.
The report added that these images facilitate the generation of realistic explicit content involving fake children and the transformation of clothed teens into nude imagery, urging immediate action from the companies behind such tools.
The Stanford Internet Observatory’s investigation focused on the LAION database, a massive repository of online images and captions utilized by leading AI image-makers like Stable Diffusion, to train AI models.
Before checking image-generating AI tools, researchers were of the view that the AI tools produced abusive imagery of children by combining information from adult pornography and benign photos of kids. However, the Observatory’s findings present a disturbing twist, with over 3,200 images of suspected child sexual abuse identified within the LAION dataset.
While LAION quickly responded to the report by temporarily removing its datasets, the implications of these images are far-reaching. Although they constitute a fraction of LAION’s vast index of 5.8 billion images, the Stanford group asserts that they likely influence the ability of AI tools to generate harmful outputs.
A significant player in the LAION database’s development is Stability AI, a London-based startup responsible for shaping the dataset. The report highlights that an older version of their Stable Diffusion model, introduced last year is still present in various applications, as the primary source for generating explicit imagery.
Although Stability AI claims to host only filtered versions and has taken steps to mitigate misuse, the challenge lies in the pervasiveness of the older model.
Generative AI Deployment Requires Rigorous Validation
The core of the issue traces back to the rapid deployment of many generative AI projects into the market, as per David Thiel, the chief technologist at the Stanford Internet Observatory.
Thiel points out that these projects were often made widely accessible due to intense competition in the field, without the necessary rigorous attention during the development phase.
Stanford researchers advocate for drastic measures to address the issue effectively. Recommendations include the deletion or cleaning of training sets derived from LAION-5B and making an older version of Stable Diffusion less accessible.
However, the complexity of retroactively cleaning up the data poses a considerable challenge, prompting a call for collaboration with child safety experts during the development of AI databases.
As schools and law enforcement agencies worldwide express alarm over the potential consequences of these findings, the spotlight is now on the AI industry to urgently address and rectify the harmful flaws within their technology.
Disclaimer
In line with the Trust Project guidelines, please note that the information provided on this page is not intended to be and should not be interpreted as legal, tax, investment, financial, or any other form of advice. It is important to only invest what you can afford to lose and to seek independent financial advice if you have any doubts. For further information, we suggest referring to the terms and conditions as well as the help and support pages provided by the issuer or advertiser. MetaversePost is committed to accurate, unbiased reporting, but market conditions are subject to change without notice.
About The Author
Kumar is an experienced Tech Journalist with a specialization in the dynamic intersections of AI/ML, marketing technology, and emerging fields such as crypto, blockchain, and NFTs. With over 3 years of experience in the industry, Kumar has established a proven track record in crafting compelling narratives, conducting insightful interviews, and delivering comprehensive insights. Kumar's expertise lies in producing high-impact content, including articles, reports, and research publications for prominent industry platforms. With a unique skill set that combines technical knowledge and storytelling, Kumar excels at communicating complex technological concepts to diverse audiences in a clear and engaging manner.
More articles
Kumar is an experienced Tech Journalist with a specialization in the dynamic intersections of AI/ML, marketing technology, and emerging fields such as crypto, blockchain, and NFTs. With over 3 years of experience in the industry, Kumar has established a proven track record in crafting compelling narratives, conducting insightful interviews, and delivering comprehensive insights. Kumar's expertise lies in producing high-impact content, including articles, reports, and research publications for prominent industry platforms. With a unique skill set that combines technical knowledge and storytelling, Kumar excels at communicating complex technological concepts to diverse audiences in a clear and engaging manner.
