iStockphoto
New research has revealed that secret messages can be hidden inside fake conversations generated by AI chatbots. The study in which this discovery was made was conducted in response to many governments around the world proposing legislation to detect, backdoor, or ban encrypted communications.
This type of legislation being proposed in numerous countries, the researchers explain, emphasizes “the need for alternative ways to communicate securely and covertly over open channels.” In their study, which was recently published on the pre-print server arXiv, the researchers claim to have come up with a way to offer a “viable alternative where traditional encryption is detectable and restricted.”
“Recent political agendas and actions significantly threaten user data privacy, as evidenced by multiple recent events,” the researchers wrote. “The UK government has demanded that Apple implement a backdoor to access users’ encrypted data. Similarly, the French government considered measures to allow message transmission within the framework of investigative requests. In a related development, Russia-backed hacking groups have devised techniques to compromise encrypted messaging services, including Signal, WhatsApp, and Telegram.
“These developments raise serious concerns about the future of secure communication. Given the potential scenario where public communication lacks encryption, it becomes crucial to explore alternative methods for embedding hidden information within publicly available content. This work addresses this challenge and proposes a novel approach to achieving covert communication under such constraints.”
What they propose is “a novel cryptographic embedding framework that enables covert Public Key or Symmetric Key encrypted communication over public chat channels with human-like produced texts.” These AI chatbot-generated texts can then be shared via platforms such as Facebook, WhatsApp or email without the secret messages being detected.
According to New Scientist…
LLMs work by piecing together text one word at a time, selecting by statistical analysis which one makes the most sense to use next. Raikwar and his colleagues altered this part of the AI and instructed it to embed the next character of an encrypted message at regular intervals in the generated text.
In this way, the AI picks the next word in a sentence based not only on whether it makes sense in context, but also so that the next encrypted character from the secret message appears in the correct spot in the sentence. If the AI gets to a point where it can’t place the next character without creating suspicious or contrived sentences, then it backtracks a few characters and tries again.
“By seamlessly integrating encryption with human-like text generation, our method provides an alternative for secure communication in scenarios where conventional encryption mechanisms are easily detected or restricted,” the researchers concluded.
Content shared from brobible.com.