How to know where things come from
A story a DECaDE in the telling, from the University of Surrey
We all care about the story behind the things we use – from the chocolate in our supermarket to the images and videos we see in our news feeds. Where did it come from? How was it made? Can we trust it?
That story is known as provenance.
For ten years, Professor John Collomosse and his team at the University of Surrey have been researching provenance technologies that help people answer these questions across both the physical and digital worlds.
The origins
The story starts in the UK’s National Archives, with ARCHANGEL, a UK Research and Innovation (UKRI)-funded project led by Collomosse to help protect the integrity of public records, including video records from the UK Supreme Court.
The project recorded provenance information about media on a blockchain – a shared, tamper-evident database maintained by multiple archives. The UK, Estonia, Norway, Australia and US National Archives were involved in the trial, which was widely reported at the time. The aim was to allow members of the public to verify the provenance of digital records when they were released.
“Institutions like the National Archives are the lens through which future generations will view the events of today, but we’re now in an age where people increasingly question institutions,” says Professor Collomosse. “ARCHANGEL explored a shift from institutional trust to technological trust, pioneering ideas like provenance metadata and visual fingerprinting to address issues of content authenticity.”
The project later extended into the area of news through Angel’s Wings – an open-source browser extension that looked for provenance records associated with photographs online. The work, which won an award for its use of computer vision and blockchain at the prestigious CVPR conference in 2019, helped show that the challenge was no longer limited to archives or official records.
Increasingly, the public encounters news and images not only through established institutions, but through a decentralised ecosystem of creators, platforms and audiences. The question was becoming broader – how can people understand the history of the digital content they encounter online?
Believing what you see
Professor Collomosse’s work has influenced global industry efforts around content authenticity. In 2019, he took up a concurrent role at Adobe Research, where he co-founded the Content Authenticity Initiative, an industry coalition working to make content provenance practical at internet scale.
The initiative has since grown to more than 6,000 members. The Initiative founded and advocated for the Coalition for Content Provenance and Authenticity (C2PA) open standard. In simple terms, C2PA records provenance information inside digital content, so that the history of a file can travel with it. This lets users cryptographically sign their photographs or videos to prove they are authentic and also disclose if and where AI was used in the creation of content.
“Many people use AI to create and tell authentic stories. Simply labelling or detecting content as ‘AI’ tells us less and less – even basic image operations such as upscaling and filling use AI today. That is why provenance is so important. Instead of trying to classify content simply as real or AI, provenance describes how content was made, edited and shared. It gives people the context they need to make up their own minds.”
Collomosse and his team have been closely involved in the standard’s technical development from the outset, chairing task forces on topics such as watermarking and blockchain.
His research team have focused particularly on creating visual watermarking and fingerprinting technologies that help these disclosures stick, even when content platforms, such as social media, strip them away.
They have also been exploring uses for provenance that address AI risks beyond fake news. Knowing the provenance of content can help creators secure attribution for their work and potentially support compensation when their work is reused for AI training.
“Once you can establish the origins and ownership of content through provenance,” Collomosse explains, “you can figure out who to credit, or even who to pay, when you use that data – including for AI training.”
Photo by Untoldstories on Pexels
Photo by Untoldstories on Pexels
Photo by Gamze Özdemir on Pexels
Photo by Gamze Özdemir on Pexels
A decade of impact
This wider agenda has been taken forward through the UKRI-funded DECaDE: Centre for the Decentralized Digital Economy – a research centre founded and directed by Collomosse at Surrey. DECaDE explores how emerging technologies can improve transparency, agency and fair value exchange across digital supply chains.
Its work builds on earlier UKRI-funded provenance projects at Surrey, such as CoMEHeRe, which examined how decentralised technologies could support the responsible monetisation of wearable healthcare data, while giving people greater visibility and control over how their data is used. Another example is the TAPESTRY project, which explored the provenance of digital identity. Both projects used decentralised technologies to help protect data integrity and support more transparent forms of data exchange.
Within DECaDE, provenance has become a foundation for rights-aware digital infrastructure. If content can carry reliable information about its origin, ownership and history, then it becomes possible to ask more ambitious questions – who should be credited when it is reused, what permissions apply, and how might value flow back to creators?
These questions now sit at the centre of national debate around AI and the creative industries. DECaDE’s work has helped inform that debate through the Time to ACCCT report with the CoSTAR National Lab, which made a series of technical and policy recommendations around provenance, creator consent and compensation. This and other DECaDE provenance work has been cited in policy settings, including the recent inquiry into copyright and AI by the House of Lords, helping explain how provenance could give the UK an advantage in building a trusted AI data economy.
Beyond the digital realm
DECaDE is a uniquely cross-disciplinary research centre, bringing together computer scientists, designers, legal scholars, and experts in digital transformation from Surrey Business School. As the work has developed, many of the design patterns around digital provenance have translated into other applications, including international trade.
“Trade relies on trust. When documents are paper-based, fragmented or difficult to verify, that creates cost, delay and risk,” Professor Glenn Parry (pictured), Co-Director of DECaDE and Professor of Digital Transformation at Surrey, explains. “We have been exploring how physical biomarkers in foodstuff, including chocolate, wine, as well as minerals, can be linked to the digital provenance record. We have also explored how those records could support the transition to electronic trade documents.”
One example is the Reduced Friction in International Trade (RFIT) project, where DECaDE researcher Mike Brookbanks worked closely with government and industry partners, including the UK Cabinet Office, and contributed to policy thinking around the move from paper-based trade documentation towards electronic trade documents.
Beyond the digital realm
After a decade of work on the subject, some students have jokingly called Collomosse “Professor Provenance”. But the nickname points to something serious – provenance has become a defining theme of Surrey’s research into trust, AI and the digital economy. Whether the question concerns a news image, a creative work, a healthcare dataset or goods moving across a border, the underlying challenge is often the same, even though the problem domains may look unconnected – how do we know where something came from, how it has changed, who has handled it, and whether we can trust the record?
DECaDE’s first funded phase is wrapping up, marked by a research showcase at the Royal Society, but the work at Surrey continues. New UKRI BRAID-funded work will continue collaboration with the CoSTAR National Lab, looking at provenance and questions of personality rights and likeness. There are also new collaborations with the BBC and the University of Oxford exploring how the provenance of narratives can be traced online using AI tools to support fact-checkers.
As generative AI, online platforms and global supply chains continue to reshape society, questions around provenance will only become more important. Surrey’s research has helped establish provenance as a foundation for trust – from the archive to the news feed, from creative rights to AI training data, and from digital content to the physical goods we use every day.
Built with Shorthand