AI researchers delete over 2,000 web links to suspected child sexual abuse imagery

Artificial Intelligence (AI) researchers said on Friday (Aug 30) that they deleted over 2,000 web linksto suspected child sexual abuse imagery from a dataset used to train popular AI image-generator tools.

A report by the news agency Associated Press early Saturday said that the dataset in the picture is theLarge-scale Artificial Intelligence Open Network (LAION).

LAIONis a huge index of online images and captions that’s been a source for leading AI image-makers such as Stable Diffusion and Midjourney.

A Stanford report which revealed the problem

In December 2023, a report by the Stanford Internet Observatory said thatLAIONcontainslinks to sexually explicit images of children, contributing to the ease with which some AI tools have been able to produce photorealistic deepfakes that depict children.

This report promptedLAION to immediately remove its dataset.

Eight months (after the report was released),LAION said that it was working with the Stanford Internet Observatory and anti-abuse organisations in Canada and the United Kingdom (UK)to fix the problem and release a cleaned-up dataset for future AI research.

Stanford commends LAION for working to fix problem but...

The Associated Press report said that Stanfordcommended LAION for significant improvements but said that the next step was to withdraw from distributing the“tainted models” that were still able to produce child abuse imagery.

The cleaned-up version of the LAION dataset comes as governments around the world are taking a closer look at how some tech tools are being used to make or distribute illegal images of children.

(With inputs from agencies)