Google's subsidiary launches recognition service for photoshopped images


02/05/2020

Alphabet subsidiary has released a service that helps to identify photoshopped pictures. The project, as the developer hopes, will help reporters identify fake information.



psyberartist
Jigsaw, a subsidiary of Alphabet, has introduced an interactive platform for identifying coordinated disinformation campaigns and the Assembler service that can identify edited photos even if they were created using artificial intelligence, reports The New York Times.

Assembler verifies authenticity of the photographs and shows where they could have been changed. The application has seven detectors, each of which is created to determine a specific photo manipulation. Five detectors capable of detecting color anomalies, photo montage, object cloning, and used camera models were developed by the University of California at Berkeley and the University of Naples in Italy. The other two identify deepfakes created using artificial intelligence. This is a proprietary Jigsaw development.

Jigsaw has also developed a platform to help identify coordinated actions in online disinformation campaigns, The New York Times writes. The database provides information about people behind the attack, their tactics and how the lies spread on social networks. Jigsaw, together with the Atlantic Council research lab Digital Forensic Research Lab, has already compiled a list of approximately 60 cases of misinformation based on more than 700 investigations published by the lab over the past five years. Emerson Brooking, a laboratory researcher, said the goal was not to create an encyclopedic list of disinformation campaigns, but to lay the foundation for a “common language” that could help the media and stakeholders learn about disinformation attacks.

source: forbes.com