Member-only story
Deepfakes are here, how do we identify them?
In this post, I will demonstrate how we can use data extracted from an image to help identify deepfakes during a crime investigation. The process may seem difficult, but with the right tools, it’s actually quite simple.
I will be using a command line on my Linux Ubuntu machine, but the same can be achieved on Windows or with online tools.
To follow along, you will need a terminal open on a Linux distro, and some images, ideally from your phone. Keep in mind that social media platforms often delete EXIF data from images by default, so it’s best to use images that haven’t been uploaded to social media.
It’s easy to extract information from an image using the command line tool like exiftool. To install it, simply enter:
apt install exiftool # install exifexiftool /path/to/image # outputs meta information of the image
I prefer to use ExifTool rather than EXIF since the latter usually outputs an error “No EXIF data available” while ExifTool still spits out at least some information.
For the sake of completeness, here is a short snippet of a random image meta information, but the list of all information can be much longer:
