Can I trust that the radish I’m eating really is organically grown on a small farm? Is the champagne I paid for through the nose actually genuine? Am I really watching the president’s interview? Authenticity always matters. This far, images and videos have been a way to ensure authenticity. In the era of AI, however, anyone can produce visual material of almost anything, artificially, without a trip to a farm or joining a top politician audience.
At VTT, we are currently running several projects related to authenticity, harnessing technology for it in the current ‘post-truth era’. In the following, three examples:
- Counterfeiting challenges everyone who produces something valuable. Ato mitigate this, as far back as about 20 years ago, we were presented with highly adhesive labels with air bubbles added in the plastic during manufacturing. The labels were all unique – differentiated by where each bubble happened to settle – and thus difficult to forge. They enabled tracing the product from the factory to the consumer. Currently, such labels would be referred to as PUF, meaning Physical Unclonable Function. When a limited amount of products is sold for a short period of time, a unique code can also be printed on each package. But what about mass produced goods, and products warehoused for longer periods? Our team of applied cryptography has studied, for example, the use of physical properties of product packaging in verifying authenticity. As an example I could mention using the unique and inherent fibre structure of package cardboard for identification.
- The journey of food from a farm to a fork is also interesting. Consumers are ready to pay extra for pure ingredients, especially if guarantees of their origin can be proved. It is estimated that in the EU area food fraud causes costs amounting to EUR 8–12 billion. The WATSON project, launched in spring 2023, is developing a holistic traceability framework and technological tools for traceability and verifying authenticity. The project focuses on five product cases: wine, honey, olive oil, meat, dairy and fish.
- In the field of verifying the authenticity of videos and images, we are currently playing a cat-and-mouse game. As new means of distinguishing the genuine from the false are developed, those producing fake images only pick up their pace. Watermarks and electronic fingerprints in images or videos help a bit. But even they are useless if the party that produced the material, such as a photographer or videographer, is not honest. We are currently developing solutions that would enable the person being filmed to ensure that their photos or videos are received as they were intended to be received.
There are several aspects related to verifying authenticity, such as copyrights and identity management issues. In the digital world, it is difficult to verify and enforce copyrights. The revenues of artists are under constant threat as illegal copies of the pieces of work created by them spread uncontrollably.
The post-truth era may sound dramatic, but a lot of interesting research projects on the matter are under way both at VTT and internationally. It is just a question of time until we will again get a firmer grip on the real thing – whether it means pixels on the screen or soup in a bowl.
Get to know our experts
Ville Ollikainen works as Research Team Leader at VTT Applied cryptography, as part of the Safe and Connected Society research area. He has graduated as Master of Science in Engineering from the Master’s Programme in Applied Physics. Ollikainen has been working at VTT since 1999. Until 2010s, he focused on media technologies research. His current team studies not only methods to verify authenticity of products and contents but also such matters as quantum-safe cryptography, the impact of quantum computing on cryptography, quantum key distribution (QKD) and management of digital identities.
