Did you know that digital images can contain secret messages that the human eye cannot see? It’s not limited to only images, you can hide messages with audio or even a video. This is known as steganography where the secret message is embedded in the binary bits of the digital medium. This is done by changing a couple of bits every single byte. Changing a few bits every byte on a digital image barely affects the color in the digital image; the human eye cannot see the difference between a regular image and a “corrupted” image. In order to “see” the message, you need to look at the digital image’s binary code and decipher it from there.

My interaction with ChatGPT 3.5:

Prompt: “Can you explain how steganography is done?”

ChatGPT: “Steganography is the practice of concealing a message or information within another non-secret file or message. The goal of steganography is to hide the existence of the message, as opposed to cryptography, where the goal is to make the message unreadable to unauthorized parties. Here’s a basic explanation of how steganography is done:

  1. Selecting a Cover Medium: The first step is to choose a medium in which to hide the secret message. This could be an image file, an audio file, a video file, or even text. The chosen medium is often referred to as the “cover medium.”

  2. Embedding the Message: Once the cover medium is selected, the next step is to embed the secret message within it. This can be done in various ways depending on the medium. For example:

    • In image steganography, the least significant bits of the pixels in the image can be modified to encode the secret message. This modification is usually imperceptible to the human eye.
    • In audio steganography, the waveform of the audio can be manipulated to encode the secret message. This can involve modifying certain frequencies or amplitudes.
    • In text steganography, the layout, formatting, or specific word choices can be used to convey the secret message without arousing suspicion.
  3. Sending or Publishing the Cover Medium: Once the secret message is embedded in the cover medium, the resulting file can be sent to the intended recipient or published publicly. To anyone who doesn’t know the message is hidden within the cover medium, it appears as a normal file.

  4. Extracting the Message: The recipient who knows to look for the hidden message can extract it from the cover medium using the appropriate steganographic technique. This typically involves reversing the process used to embed the message, such as extracting the modified bits from an image or analyzing the audio waveform for hidden data.

Steganography can be a useful tool for covert communication, but it’s important to note that it’s not inherently secure. If someone suspects that steganography is being used, they can analyze the cover medium for hidden data or use steganalysis techniques to detect the presence of a hidden message. Therefore, steganography is often used in conjunction with encryption and other security measures to provide a higher level of secrecy.”

Examination of Chat GPT 3.5

Similar to what we did in class, when ChatGPT says the least significant bits of pixel modified, he means changing either the last bit or the last two bits of a byte (8 bits) to encode the secret message. Changing the least significant bits has minimal effect on the image so anyone looking at the image for the first time won’t see anything suspicious.

I’d say my work and understanding is 70% me and 30% Chat GPT because of my background in digital systems.