Google Develops AI Software to Combat Child Abuse Images
Google has developed a new software that uses artificial neural networks to identify child abuse images on the internet. The technology, known as AI, is far superior to humans in terms of speed and accuracy. The software can identify suspicious images and videos without the need for them to be reported, and it categorizes them into priority groups. The marked images and videos are then presented to human moderators for review. The new software has been tested and found to be able to check 700% more suspicious files than before.
Previously, images or videos of child abuse had to be reported by other users and then reviewed by moderators. Once marked, the images were stored in a database where they could be recognized by their digital fingerprints if they were uploaded again. This method, known as PhotoDNA, is currently the most effective way to combat child abuse images. The software is used by Facebook, Twitter, Google, and Microsoft. However, the process of reviewing reported images is time-consuming and emotionally taxing for moderators. The new software is expected to ease the burden on moderators and law enforcement agencies.
The new Google software could also help to reduce the spread of child abuse images in the darknet, where an estimated 80% of all data is related to child abuse. The darknet operates differently from the regular internet, making it difficult to locate and delete such content. The new software is expected to be made available to other platforms for free.
