Over last couple of years, fake information has been a constant topic of discussion among internet users following Facebook’s Cambridge Analytica scandal. In order to fight against the spread of false information, Adobe has introduced several tools to detect manipulated images.
In an official blog post, Adobe revealed that the company worked with scientists from US Berkeley to create a neural network which can detect fake photos and work to restore the original images.
While Adobe already has features to prevent people from altering certain images such as bank notes, the new tool will allow users to know if the images they are seeing was altered beforehand. In order to find out if the new tools were working properly, the company asked humans if they can detect manipulated images. The results showed that humans were able to spot edited images only 53 percent of the time. On the other hand, the AI detected manipulated images with 99 percent accuracy. Moreover, the tool highlights the areas of the images which it thinks was edited using Adobe Photoshop. As a result, users can take a closer look to determine if any information was altered from the original images.
Considering the spread of fake news over the internet, the latest tool has potential to make huge impact. However, given that there are so many ways people can edit images, it’s unlikely the tool will result in massive reduction of fake news. Despite that, the AI shows that it’s possible to create AI which could help us to combat the spread of fake news in the coming years.