Security design is truly an art of deception. It requires a special skill to create something beautiful and symbolic, but it also demands a high level of ingenuity and creativity to perfect. In many ways security designers are more like engineers or architects than artists. From conception to production, precision in the details is key. To be able to harmonise aesthetics, security features and special materials or inks across multiple print processes into a unified mass-produced document is quite a feat. Additionally, all of this is achieved in a unique design that dynamically changes as you interact with it and alter your perspective from macro to micro as well as the invisible spectrums. So, with such a task, where do you even begin?

We will start by asking some questions. What is the secure document going to be used for? Is it a national symbol, such as a banknote or passport, or a less prominent, but equally important document such as a birth certificate or tax stamp? Who needs to be able to verify the authenticity – highly trained inspectors or the general public, or perhaps both? How sophisticated are the counterfeiters? Are they ordinary people with simple office equipment or are we dealing with wealthy and resourceful organised crime or even governments? Of course, you also have to ask: how much time is there to produce and what is the budget? Finally, how long should this document withstand the test of time – in terms of durability, security and ever-evolving technologies? All of these questions are extremely important to consider before you even start, because you must begin with the end in mind.

Securing the digital future
Technologies are constantly changing. We have to think about how powerful equipment such as scanners, printers and smartphones will be in the future. Many past technologies that were considered strong and ‘copy-proof’ have since become less effective due to the readily available imaging equipment. Nowadays, you can buy all-in-one scanners and printers with a very high resolution from any office store for less than 100 US dollar. Think about how this has changed the counterfeiting and anti-counterfeiting game. At the same time, our world has become much more con­nected and monitored in this digital age we are in. So, while it may seem easier to counterfeit in some ways, it is much harder to stay under the radar and not get caught eventually. This begs the question: should secure documents and anti-counterfeiting techniques become more digital?

Obviously, in every sense, there are positives and negatives to becoming more digital. In terms of the negatives, a main concern for anything digital is and will likely always be the threat of cyberattacks such as hacking and the implications for robustness and redundancy when things like ‘bugs’, glitches or simply dead batteries or power outages occur. Now, since my background is in graphic security features such as Scrambled Indicia® and all of the advanced applica­tions of it, I will provide a positive case for going digital from that perspective.

The digital advantage
Prior to the advancements in imaging equipment, many optical security indicias, void pantographs and water­marks were extremely difficult to copy with any degree of precision. In the early days of film and camera-based image generation, it required a high degree of expertise to even create images like these, let alone to counter­feit them. There was a time when anything higher than 150 LPI (lines per inch) was considered ‘copy-proof’. The new standard tends to be over double that as a bare minimum and even that can sometimes be copied with the right equipment and know-how.

So, what does this mean for features like these; is it game over? No, not at all! In fact, some of them are stronger than ever. This is where the digital approach really shines – both in terms of generation and authentication. On the generation side of things, variable printing and features are the key to securing lower frequency images. You might be able to copy them, depending on how they are applied but they prevent easy alteration. When designed by incorpo­rating highly unique, ever-changing data into these optical features, they can be very effective. Imagine having multiple variable security features throughout virtually every aspect of a document, which contain unique data. It would require a great deal of customisation to reproduce and would therefore act as a deterrent to counterfeiters.

In identity documents, you can personalise security within the main and secondary or ‘ghost’ photos, the background design, invisible print, perforations, laser engravings/ablations and even foils. Other documents such as birth certificates, money orders or tax stamps can be layered in the same way, containing dates, financial figures, serial numbers, etc. Counterfeiters would have to hit constantly moving targets. Incorporate these layers of different variable features, printed or imaged across multiple processes, with multiple inks and materials and if you are a counterfeiter, it is time to move onto an easier document!

Because of this strength when it comes to variable security features, I predict we will see a lot more digital printing emerging as truly secure in nature. Traditionally, this is a frowned-upon approach. Various materials and processes such as offset/litho and intaglio printing are considered more secure, which they typically are due to quality and restricted access. However, when effectively taking advantage of the uniqueness and variability that digital processes have to offer, it can be equally challenging to counterfeit. You can also then use hardware and software access control keys to restrict unauthorised use, as opposed to using physical chain of custody and inventory control to protect supplies of special inks or materials, for example.

Smartphone authentication
Now, going into the authentication benefits of a digital solution, there are many. There are also numerous ways to approach digital authentication, in particular which algorithms are used. When doing smartphone authen­tication, there is a lot to take into consideration but at the end of the day, the majority is being done through the camera interface, and the core algorithms are going to be image processing-based. There are some excep­tions: NFC (Near Field Communication) for example is more of an identification or data storage feature than a security feature, and can be verified without the use of the camera. However, in most cases, whether it be a printed security image, optically variable ink, hologram or anything micro, the analysis of the feature will be based on images captured by the camera.

You must have noticed that practically every security company has launched some sort of ‘smartphone authentication’ application, trying to get on this band­wagon. Suffice to say that if an app just acts as a QR code scanner or merely shows a diagram of how to visually authenticate a sample, then it is not doing any actual authentication. Apps like these take away from the seriousness and actual utility of a smartphone for true authentication. It will drive development if we pursue this goal as an industry and hold authentication apps to a high standard. Another thing to remember is that any serialisation is really suited best for chain of custody, distribution or track and trace applications –
in other words: identification, not authentication.

Overcoming the challenges of smartphone authentication
Using a smartphone to authenticate security features is incredibly challenging. You have to account for variations in printing, materials, smartphone hardware, software, camera quality, lighting conditions and user skill, to name but a few. On the hardware side, the iPhone platform alone has over 15 models, and Android has literally thousands throughout the world. Software updates and firmware variations are even more diver­sified, and the way each model captures, processes and saves images is all over the map. Print quality, users and environment further complicates things. There are essentially infinite variations you have to take into consideration: all of the above as well as different print processes, ink density, matte vs. gloss substrates, shaky hands, different angles of scanning, skewing, distance, pitch, roll, and of course lighting.
In terms of analysing a printed security feature, there is a wide range of original quality that happens on press. In some cases, the acceptable quality range allows for samples that are bad enough they start to look like good counterfeits. And likewise, on the opposite spectrum, there might be quite a range of quality in the counterfeits as well. You can’t simply come up with a one-dimensional algorithm that looks at ‘black and white’ criteria, and this leaves a margin for error that is unacceptable when performing security analysis.

Machine learning
So how can you possibly account for such variation in quality, both in originals and counterfeits, while performing reliably? One approach gaining traction is artificial intelligence, or ‘AI’ for short. Not in terms of android inspectors, but rather what is behind actual AI: machine learning. If you haven’t heard of this, you certainly will. It already impacts your life and you probably don’t even know it. Soon it will be behind virtually every aspect of travel, commerce, com­plex data crunching and yes, security. With ma­chine learning it is now possible to consistently spot even the best counterfeits with just a smartphone.

The benefits of machine learning are abundant when it comes to digital authentication and ana­lysing physical security. I have found that features such as Scrambled Indicia®, originally a purely lens-based optical tech­nology that relied on human interpretation, has become exponentially more secure when digitally verified. In virtually all cases, an ultra-high-resolution scan and reprint in offset/litho, which cannot be dis­cerned from the original copy by a human eye, is very accurately and reliably identified by the digital algorithm based on machine learning. This algorithm spots the counter­feits that I, as an expert, could not spot myself, not one single time. It is quite amazing, actually, but what is almost more impressive, is how robust it is. This process can handle a very wide range of press variation and original quality parameters, which is extremely important for quality control. The results identify a variety of counterfeit attempts, ranging from simple colour copies to full-blown sophisticated reproductions on actual presses. I have never seen anything like it, but I guarantee you will be seeing more of it.

Summary
Despite this seemingly impossible goal of reliable smartphone authentication, the outcome is absolutely necessary to secure documents in our future. Counter­feiters are just too good at what they do. They are only going to get better too, and we need to up the security game intensely to overcome this threat. We need highly secure, ever-evolving features that push the limits, while remaining accessible for public verifi­cation. The only problem is that there is an inverse relationship between security level and ease of authentication. This means that the more secure something is, the more difficult it generally is to verify and vice versa. How can you equip virtually anyone without training to verify multiple virtually forensic-level security features in one easy step? You need an incredibly powerful and intuitive device, which we all happen to own and carry everywhere we go. That is why it is time to take smartphone-based authentication seriously and explore the capabilities. As security designers and printers liaise with software engineers to achieve this common goal, the future of authenti­cation can only get stronger and help us stay ahead of the counterfeiters.