Introducing Trusted Media Capture
Verifying the authenticity of photos and videos is already difficult, but will become near impossible with the advent of computer generated media. We present a new technological approach, now funded by Google through its DNI Fund.
In its July article titled ‘How “fake news” could get even worse’ The Economist described what those of us working in machine learning, computer vision and adjacent fields have known for a while now — computers are on the verge of being able to generate credible video and audio of anyone saying anything, and highlighted the fact that these will be equally good in fooling both humans and machines. For us at Serelay this suggests we need to rethink not just how we verify digital media, but far more importantly — how we capture it. Why? — it allows us to solve the problem at the source using the technical infrastructure we already (collectively) own. Let me explain:
The vast majority of media today is captured on mobile devices. We know these devices have increasingly higher spec cameras and screens , but there is something else we tend to forget in the context of media capture: they also have sophisticated operating systems embedded with a relatively robust security infrastructure and a large array of sensors.
These are robust enough for me to pick up my phone and immediately transfer a $10,000 from my company bank account to a 3rd party recipient. My bank would have the confidence to know the transaction originated from me and that its content is authentic. Why then, can’t I send a photo I captured on my device to Facebook, CNN, or anyone else for that matter, giving them the same level of confidence pertaining to it it’s authenticity and origin? Well, we argue you can.
By applying similar security protocols to those used in financial transactions and utilising the full breadth of sensor metadata available on a mobile device we aim to create what we call: ‘trusted media capture’.
‘trusted media capture’ is founded upon 3 principles:
- WYSIWYG (What You See Is What You get)
the photo or video you’re watching has not been edited since capture. It provides an authentic, unedited representation. - Physical Barrier to Entry
The person who captured the media was present at a specific location at a specific time - Automatic for the People
The verification process does not require a manual and/or expert component
(for the techies among you — you’re right, it’s not that simple… rooted devices, jamming GPS signals or other sensors, taking a photo of a doctored photo… we know… and we’re working on it. If you want to join us, let us know at info@serelay.com…)
Oh, and it’s not just fake news either — whether we’re swiping right on a tinder profile or booking an AirBnb, our ability to trust digital photos and videos plays a pivotal role. Badoo, a dating service with over 300 million registered users launched last year a photo-verification feature. The fact it needs to use 5000 (!) human moderators to do this, underscores our argument for trusted media capture as a technology whose time is now long overdue. (similarly, AirBnB would only list a photo as ‘verified’ if it was taken by an AirBnb commissioned photographer.
Over the next few months we’ll be rolling out our trusted media capture technology in collaboration with a handful of early adopter platforms. For us, these are exciting times, and we’d love to talk to you if you’re an organisation concerned with verifying digital news, dating profiles, classified listings, or anything else. Feel free to drop us a line at — info@serelay.com
Serelay’s ‘trusted media capture’ tech is now funded by Google through its DNI Fund
Thanks for making it all the way here…
If you made it so far, please follow us at @rozoulay and/or @serelayTech for updates.
