Privacy Violated: AI Companies Accused of Using Children's Photos Without Consent in Brazil

Privacy is dear to every single human, and when it gets violated, serious human rights issues arise, and people are forced to take legal action against the aggressor. But when it comes to the privacy leak of children, things get very intense and serious because no one wants to see an innocent child getting harmed in any way.

Shock waves have been sent across Brazilian families due to AI companies using children’s pictures without their parent's permission to train AI models. The images are taken from the internet social media platforms to train AI in large datasets. The biggest fear that people have is that these pictures can be used to create deepfakes that can bring massive harm to the welfare and mental health of the coming small minds.

The public has cried out their concerns to the government over this illegal use of their kids' pictures for advanced AI learning, but no strict action has been taken yet. According to Hye Jung, a Human Rights Watch activist, it is the LAION-5B dataset that carries links to detectible pictures of Brazilian kids and can be traced with ease. What's more terrifying is that most of the pictures have the names of kids too in the URL where the image was stored along with the other information about the picture such as when and where it was taken.

As per HRW (Human Rights Watch), the pictures of children cover their entire childhood, including the moments of their birth, age celebrations, and random home pictures of doing activities. LAION-5B being the biggest dataset available in the world since March of 2022, is built by using millions and billions of images taken randomly from the Internet. Furthermore, a research publication issued that they have found several thousand images that carry child abuse and sexual content.

The allegations made by Stanford Internet Observatory about the Rated R content about child abuse were responded to by the LAION through the statement that they have a zero-tolerance policy for illegal content. To make amends, they even took the LAION-5B model offline.

Although there are strict laws on the use of AI to create sexual images, no legal laws or actions have been set so far by the government to prevent the scrapping of children’s images off the internet.

Image: DIW-Aigen

Read next: YouTube Equips Its Search Bars On Android Devices With Google Lens
Previous Post Next Post