Infotech How Photoshop will fight (very concretely) against “fake news”

How Photoshop will fight (very concretely) against “fake news”

Since mid-July, torrential rains have fallen in Sudan, causing at least 124 deaths. However, the photos used to illustrate these devastating floods have no relation to the situation on the spot: they sometimes go back several years and were taken, jumbled, in India, Haiti, Somalia or New Zealand. Orleans, reports AFP. Another case: a video shows two men fighting in a parking lot, certifying that it is a Guinean minister and his chief of staff. However, as the agency deciphers, this fight took place in 2016 between a professor and a student.

These two recent examples show the omnipresence of manipulated and false internet publications, which nevertheless accumulate tens of thousands of shares on social networks. Problem summarized by the English expression “fake news”. And that’s exactly what the multinational Adobe intends to deal with, best known for its Photoshop, Premiere, Illustrator creation software and its PDF document format.

Online disinformation is one of the biggest challenges today and we all have a responsibility : camera manufacturers, image software publishers, content distributors, slice William Allen, vice president of Adobe.

And to continue: Faced with false information, of course, it takes media education to take a step back from what is posted, while working with AI tools. [intelligence artificielle] to better detect if things have been changed. We believe that there is a third aspect which is to attribute more precisely the contents. And on this point, Adobe intends to deploy a new free standard, under the name of Content Authenticity Initiative (CAI). It will be launched this Tuesday, October 20 in partnership with a wide variety of players: Twitter, Microsoft, chipmaker Qualcomm, the American daily “New York Times”, the American magazine “Inc.”, the British radio BBC , the Canadian channel CBC, the NGO Witness …

“A layer of trust, transparency”

Concretely, the CAI takes the form of an information tab added first to images, then to videos and sounds, and one day to virtual reality (VR) and augmented reality (AR) images. This tab allows you to identify at a glance the author (s), the dates and places of creation, and any modifications (retouching, cropping, etc.). In short, a kind of new metadata (named after the technical data attached to the photos), encrypted at the heart of the content and impossible to modify.

We want to restore confidence in what we see online, continues William Allen. When we see a photo, watch a video or listen to a sound, we will be able with one click to know with certainty who created it, where it was captured, how it was edited. It’s a layer of trust, transparency, to better understand, especially in the face of very viral content on social networks.

Completely open and free, the tool will be integrated into Adobe software, starting with Photoshop. And Americans are hoping that camera and smartphone makers, software companies, internet browsers, social media, media editorial staff, etc. will use it to provide additional information to the public.

We are working to make it a standard accessible to everyone, not just Adobe customers, insists Andy Parsons, director of the Content Authenticity Initiative. This is a completely open source framework, and it is hoped that it will be implemented quickly everywhere, in all phones, in all networks, in all media, etc. With the hope of countering the proliferation of fakes news and others deepfakes (these videos that make anyone do and say anything).

Notable side effect: content creators (photographers, videographers, etc.) will be better credited for their productions and will be able to assert their rights more easily. It remains to be seen how long it will take for the CAI to be invoked in a trial.

Leave a Reply

Your email address will not be published. Required fields are marked *