Computer program to eavesdrop on hidden messages has been developed

By Mark Waghorn via SWNS

A computer program that makes it possible to eavesdrop on hidden messages has been developed by scientists.

It opens the door to perfectly secure digital communications - combating terrorism and organized crime.

The algorithm can embed information in photographs and even music - concealing content in a way that cannot be detected.

It may also be widely used in social media and private messaging - helping vulnerable individuals such as dissidents, investigative journalists and aid workers.

The team from the University of Oxford in the UK employed a technique called steganography - the electronic equivalent of invisible ink.

An example would be a Shakespeare poem being placed inside an image of a cat.

Lead author Dr. Christian Schroeder de Witt, from Oxford, said: "Our method can be applied to any software that automatically generates content, for instance, probabilistic video filters, or meme generators.

"This could be very valuable, for instance, for journalists and aid workers in countries where the act of encryption is illegal.

"However, users still need to exercise precaution as any encryption technique may be vulnerable to side-channel attacks such as detecting a steganography app on the user's phone."

AI-generated content is increasingly used in ordinary human communications - fueled by products such as ChatGPT, Snapchat AI-stickers and TikTok video filters.

As a result, steganography may become more widespread as its presence will cease to arouse suspicion.

Communication breaches can be disastrous on multiple levels - from causing irreversible reputational damage to a business to putting lives at risk.

In 2021 private photos, personal chats, sensitive data and the locations of over 900 million active Apple users were compromised.

Each day billions of messages are sent over the internet - some containing very sensitive information. Much effort goes into making sure they are unreadable for anyone bar the intended recipients.

Steganography differs from cryptography by hiding data in plain sight rather than scrambling it.

It has been likened to the classic conjuring trick of misdirection. Something that looks like nonsense is quite likely to be of interest.

Co-lead author Samuel Sokota, a Ph.D. student at Carnegie Mellon, Pittsburg, said: "The main contribution of the work is showing a deep connection between a problem called minimum entropy coupling and perfectly secure steganography. By leveraging this connection, we introduce a new family of steganography algorithms that have perfect security guarantees."

Governments and the military want to pass messages securely. No one can analyze every piece of data that could potentially be captured.

If a piece of data looks like, say, a picture but actually it contains a secret hidden document, no one will know to conduct further analysis.

Steganography is already in used in everyday life to "digitally watermark" electronic data with information such as the copyright owner.

Dr de Witt and colleagues tweaked the system by adopting a more effective mathematical formula that makes innocuous and sensitive content statistically the same.

Tests on an open-source language model and a text-to-speech converter found the algorithm was perfectly secure.

It also achieved up to 40 percent better results than other steganography methods across a variety of applications - enabling more information to be concealed.

The study in arXiv offers hope of improved capacity for data compression and storage, too.

Decoding encrypted messages has long been studied, with efforts such as those at Bletchley Park during World War II being rightly celebrated.

Detection of hidden messages - known as steganalysis - has no such pedigree.

The researchers have filed a patent for the algorithm, but intend to issue it under a free license to third parties for non-commercial responsible use by academics, humanitarians and trusted third parties.

Co-author Professor Jakob Foerster, from Oxford University, said: "This paper is a great example of research into the foundations of machine learning that leads to breakthrough discoveries for crucial application areas. It is wonderful to see that Oxford, and our young lab in particular, is at the forefront of it all."