Mark the good stuff: Content provenance and the fight against disinformation
Posted by Charlie Halford on , last updated
±«Óãtv News’s Verify team is a dedicated group of 60 journalists who fact-check, verify video, counter disinformation, analyse data and - crucially – explain complex stories in the pursuit of truth. On Monday, March 4th, Verify published their first article using a new open media provenance technology called C2PA. The C2PA standard is a technology that records digitally signed information about the provenance of imagery, video and audio – information (or signals) that shows where a piece of media has come from and how it’s been edited. Like an audit trail or a history, these signals are called ‘content credentials’.
Content credentials can be used to help audiences distinguish between authentic, trustworthy media and content that has been faked, by showing them where it has come from. Ultimately, they are the ones who decide whether they trust that media. The digital signature attached to the provenance information ensures that when the media is “validated”, the person or computer reading the image can be sure that it came from the ±«Óãtv (or any other source with its own ).
This is important for two reasons. First, it gives publishers like the ±«Óãtv the ability to share transparently with our audiences what we do every day to deliver great journalism. It also allows us to mark content that is shared across third party platforms (like Facebook) so audiences can trust that when they see a piece of ±«Óãtv content it does in fact come from the ±«Óãtv.
For the past three years, ±«Óãtv R&D has been an active partner in the development of . It has been developed in collaboration with major media and technology partners, including Microsoft, the New York Times and Adobe. Membership in C2PA is growing to include organisations from all over the world, from established hardware manufacturers like Canon, to technology leaders like , fellow media organisations like NHK, and even the Publicis Group covering the advertising industry. Google has now joined the C2PA steering committee and social media companies are leaning in too: they are actively assessing implementing C2PA across their platforms.
As more organisations sign up to C2PA, people will become used to seeing images or video with content credentials included. We hope this will give more and more people the tools they need to judge the authenticity of what they’re seeing for themselves and preserve the integrity of the information ecosystem.
How we got here
At the ±«Óãtv, we pride ourselves on being the world’s most trusted news media organisation. We believe that ‘If you know how it’s made, you can trust what it says - trust is earned.’ We are committed to delivering impartial, accurate news to our audiences. This matters now more than ever as we face a growing onslaught of fake or misleading information.
±«Óãtv R&D has been thinking about this problem for a long time and, back in 2019, we decided to do something about it. In collaboration with Microsoft, the New York Times and CBC/Radio-Canada, ±«Óãtv R&D began exploring some early technology solutions to the problem of disinformation. We soon began to collaborate with another like-minded organisation called the Content Authenticity Initiative, headed by Adobe.
Together, we founded the C2PA, an organisation set up to develop open media provenance standards. It represents a collection of members from across many different industries, all united by the desire to represent the provenance of media accurately and transparently. To ensure that C2PA would work for news media organisations, Project Origin held a series of workshops to assemble the requirements for this technology. We learned that news organisations needed to be able to add their own data to media so they could provide the same kind of context to their readers that a news report would. They also wanted to add their own provenance data so that users could tell where the content had come from.
We have come a long way in the past five years. Today, members of C2PA all over the world are planning to make use of the specification’s media provenance features in lots of interesting ways. This could include labelling AI generated images, presenting the original context of a video when it is shared on social media, or providing a complete list of all the edits done to an image since it was taken.
We hope that this is just the beginning. C2PA was designed to represent the complete chain of provenance of all media, which, when available, will provide users with an unparalleled amount of information to help them decide whether the content they are consuming is genuine and trustworthy.
How does it all work?
Although it may seem straightforward, publishers must orchestrate a complex process before an image or video is ready to be shared online. What does this mean in practice? Let’s take an example from ±«Óãtv News. An image is captured in the field by a photographer. The photographer may upload the picture from their camera to their computer to do some light edits. These could include improving the lighting of their subject or correcting white balance. Then, they may send the edited image to an agency. There, the agency will catalogue the image and its metadata (information about where and when it was captured, an image description, etc). The agency may also lightly edit the image – perhaps cropping to better emphasise the subject, or resizing or converting the format to whatever their customers need. Then a news organisation, like the ±«Óãtv, might use that photo. Again, depending on where it’s being used, the image might be edited again (perhaps it will be used in a homepage slot that requires a square picture).
Finally, the image is made available to the public as part of a news article.
Because , any edits to the content will break it. This is on purpose – and it is what allows C2PA signed images to be so secure. It is too difficult to allow some edits to work, and others to fail, and so we require that each edit is individually signed. This, when available, provides a chain of edits that can be viewed all the way back to the point at which the camera took the photo. Each edit, signed by the entity that made them, can then let the user know what edit was performed. The spec also allows for thumbnails at each step to be recorded as well.
All of this underscores the primary intention of C2PA: to provide secure, authenticatable provenance that allows users to make their own decisions on whether content is trustworthy. There is still some way to go before this is possible, particularly when it comes to video and audio.
We should note that while this is an important tool for giving audiences the information they need to distinguish been trustworthy and untrustworthy information, it cannot be used as a fact checking mechanism to establish whether an image itself is real or fake. It simply allows users to see transparently what has been done to an image before it has been published and armed with that information, to make decisions about the authenticity of the image themselves.
We are witnessing a turning point for the news media industry. We hope this marks the beginning of a renewed fight back against online mis and disinformation, giving news media organisations another important tool for reigniting trust in the information ecosystem.
±«Óãtv R&D - Is this real? Provenance in generative media
±«Óãtv R&D - Increasing trust in content: Media provenance and Project Origin
±«Óãtv News - Transparency tool launched by ±«Óãtv Verify