Researchers have proposed an uncrackable combination of invisible ink and artificial intelligence to protect written documents.
Encrypted messages in invisible ink may sound like something out of a spy novel, but they can serve important security purposes in real life. Even as electronic records progress, paper is still a common way of preserving data – possibly increasingly in light of the widespread state-sponsored hacking of computer systems through cyber warfare.
Conventional paper information protection is primarily based on stimuli-responsive functional materials that can display color or luminescence under external stimuli.
While “invisible ink” can hide classified economic, commercial, or military information from prying eyes, many popular inks contain toxic compounds and are easily cracked if their coding methods are predictable, such as with the use of light, heat, or chemicals.
Researchers have now developed and printed complex coded data using regular ink and a carbon nanoparticle-based invisible ink that requires both UV light and a computer that has learned the code to reveal the correct hidden message.
The ink was prepared by dissolving carbon nanoparticles in water, which has high quantum yield and excellent light stability and salt stability, ensuring the integrity of information in complex environments.
Carbon nanoparticles, which have low toxicity, can be essentially invisible under ambient light, but can create vivid images when exposed to ultraviolet (UV) light – a modern take on invisible ink.
In addition, advancements in artificial intelligence (AI) models can ensure that messages can only be deciphered on well-trained computers.
Weiwei Zhao, Kang Li, Jie Xu and colleagues trained an AI model to identify and decode symbols printed in a fluorescent carbon nano particle ink, revealing hidden messages when exposed to UV light.
The researchers made carbon nanoparticles from citric acid and cysteine, which they diluted with water to create an invisible ink that appeared blue when exposed to UV light.
The team loaded the solution into an ink cartridge and printed a series of simple symbols on paper with an inkjet printer.
The team then learned an AI model, made up of multiple algorithms, to recognize symbols illuminated by UV light and decode them using a special codebook.
A five-layer convolutional neural network (one of two mainstream architectures in today’s artificial intelligence fields) was specially trained based on the ultraviolet-generated symbols printed with invisible ink.
Finally, they tested the AI model’s ability to decode printed messages using a combination of both plain red ink and UV fluorescent ink.
With 100 percent accuracy, the AI model read the regular ink symbols such as “Stop”, but when a UV light was displayed on the writing, the invisible ink illustrated the desired “Begin” message.
Because these algorithms can detect minute changes in symbols, this approach has the potential to securely encrypt messages using hundreds of different unpredictable symbols, the researchers said.
It was possible to design unpredictable and highly complex password books to further increase information security.
This smart strategy could open new opportunities for high-quality paper information encryption and also proposes new ideas for the applications of carbon nanoparticles and artificial intelligence.
The study was funded in part by the Shenzhen Peacock Team Plan and the Bureau of Industry and Information Technology of Shenzhen through the Graphene Manufacturing Innovation Center.
The team’s findings are published in the American Chemical Society’s journal Applied Materials & Interfaces.
Looking at the world of espionage, E&T went in search of the modern spy to find out what skills a real Secret Service agent needs and to see if the “James Bond” legend matches reality.Perhaps an invisible ink pen will become a standard problem in the future.