Chatbots Can Be Used To Hide And Transmit Secret Messages Inside Fake Conversations

Researchers have made a groundbreaking discovery that secret messages can be hidden inside fake conversations generated by AI chatbots, providing a new and innovative way to communicate securely over open channels.

The study was conducted in response to governments around the world proposing legislation to detect, backdoor, or ban encrypted communications. This trend has highlighted the need for alternative methods of secure communication that can evade detection by traditional encryption mechanisms.

"Recent political agendas and actions have significantly threatened user data privacy," stated the researchers. "The UK government's demand for Apple to implement a backdoor to access users' encrypted data, as well as France's consideration of measures to allow message transmission within investigative requests, demonstrate the growing concern about secure communication."

In light of these developments, Russia-backed hacking groups have also devised techniques to compromise encrypted messaging services like Signal, WhatsApp, and Telegram.

"These developments raise serious concerns about the future of secure communication," warned the researchers. "If public communication lacks encryption, it becomes crucial to explore alternative methods for embedding hidden information within publicly available content."

A Novel Approach to Covert Communication

The researchers propose 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 be shared via platforms like Facebook, WhatsApp, or email without the secret messages being detected. The key to this approach lies in the way LLMs (Large Language Models) work by piecing together text one word at a time, selecting which word makes the most sense to use next based on statistical analysis.

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. This allows the AI to pick the next word in a sentence not only for context but also so that the next encrypted character from the secret message appears in the correct spot.

"This work addresses this challenge and proposes a viable solution to achieving covert communication under such constraints," concluded the researchers.