Skip to main content

This algorithm can hide secret messages in regular-looking text

FontCode: Embedding Information in Text Documents using Glyph Perturbation

Whether it’s hiding messages under the stamps on letters or writing in invisible ink, people have always found ingenious ways of using whatever technology they have available to write secret messages. A new project carried out by researchers at Columbia University continues this tradition by using some deep learning technology to embed encrypted messages in otherwise ordinary looking text.

Recommended Videos

Fontcode” works by making incredibly subtle modifications to everyday fonts like Times New Roman and Helvetica, embedding coded messages inside them. These changes are so subtle that the average person viewing the text would be incredibly unlikely to notice them. They include such alterations as slightly sharper curves or a minutely thicker stem on a particular letter. Each letter has 52 different variations, which makes it possible to encode both lowercase and capital letters within every letter of the alphabet, along with punctuation marks and numbers, too.

Please enable Javascript to view this content

The researchers then trained a deep learning neural network to recognize these letters and to match them back to the coded letters in the secret message. With the right smartphone app and just a short period of time for processing the data, it’s possible to decode a secret message from the document it’s embedded in. Simply aim your device at the text and, as if by magic, the real message can be extracted.

Would such a technique ever be applied in the real world? Almost certainly not in everyday conversations, where the idea of having to send one another false text documents to embed a short hidden message sounds like way too much work. However, that doesn’t mean that this is relegated to being an impractical, albeit impressive, demo. It could certainly have applications in the security field, as well as potentially as an invisible watermark. Heck, you could even use it as a to- secret QR code to link to a web address.

A paper describing the project, titled “FontCode: Embedding Information in Text Documents using Glyph Perturbation,” will be presented later this year at the Special Interest Group on Computer Graphics and Interactive Techniques (SIGGRAPH) 2018 conference.

Someone should probably forward this research on to the James Bond producers before then, though. We can totally imagine Daniel Craig using the “Fontcode” algorithm in the next 007 movie!

Luke Dormehl
Former Digital Trends Contributor
I'm a UK-based tech writer covering Cool Tech at Digital Trends. I've also written for Fast Company, Wired, the Guardian…
I’ve experienced the next era of AI, and I’m never going back
Launching Gemini Deep Research query on Chrome desktop.

Ever since ChatGPT arrived on the scene, the hype around AI has only intensified. As talk of Artificial general intelligence (AGI) and “superintelligence” — yeah, OpenAI chief, Sam Altman, is now talking about that — heats up, we have another buzzword to deal with.

Say hello to Agentic AI. In simpler terms, AI agents that are supposed to automate a chunk of our digital chores. Think of Gems in the Google lexicon. Custom GPTs by OpenAI. Or Copilot Actions by Microsoft.

Read more
Everything you need to know about AI agents and what they can do
a hollow man under light

The agentic era of artificial intelligence has arrived. Billed as "the next big thing in AI research," AI agents are capable of operating independently and without continuous, direct oversight, while collaborating with users to automate monotonous tasks. In this guide, you'll find everything you need to know about how AI agents are designed, what they can do, what they're capable of, and whether they can be trusted to act on your behalf.
What is an agentic AI?
Agentic AI is a type of generative AI model that can act autonomously, make decisions, and take actions towards complex goals without direct human intervention. These systems are able to interpret changing conditions in real-time and react accordingly, rather than rotely following predefined rules or instructions. Based on the same large language models that drive popular chatbots like ChatGPT, Claude, or Gemini, agentic AIs differ in that they use LLMs to take action on a user's behalf rather than generate content.

AutoGPT and BabyAGI are two of the earliest examples of AI agents, as they were able to solve reasonably complex queries with minimal oversight. AI agents are considered to be an early step towards achieving artificial general intelligence (AGI). In a recent blog post, OpenAI CEO Sam Altman argued that, “We are now confident we know how to build AGI as we have traditionally understood it,” and predicted, "in 2025, we may see the first AI agents ‘join the workforce’ and materially change the output of companies.”

Read more
You can now try out ChatGPT Search for free
The ChatGPT Search icon on the prompt window

As part of its "12 Days of OpenAI" event, OpenAI has yet another update for ChatGPT, this time bringing its Search feature over to the free tier. The Google Search alternative was previously only for paid subscribers in the ChatGPT Plus or Pro tiers.

"We rolled it out for paid users about two months ago," Kevin Weil, OpenAI's chief product officer, said during Monday's livestream. "I can't imagine ChatGPT without Search now. I use it so often. I'm so excited to bring it to all of you for free starting today."

Read more