In an interesting development beneficial potentially for covert activities or discreet dissident discussions, computer scientists have designed a way to surreptitiously convey classified information. The researchers claim to have developed the first-ever algorithm capable of burying messages within normal text, images or audio with absolute discretion. The method is so discreet that an external party can't even determine if a secret message is being communicated. They applied an AI program to pen innocent-looking chat messages with hidden content. For anyone externally observing the chat, all they see is normal AI-generated conversation. While they might be able to recognise the use of AI, deciphering any secretive information within the content is virtually impossible. Making strides in this field, the researchers have developed an algorithm capable of skilfully concealing a secretive message within a range of memes that are selected depending on the context. This approach does demand a significant computational effort, however, the team has utilised recent progress in the field of information theory to rapidly make near-optimal selections. The implications of the advancements in this AI-based form of steganography are significant. As AI that can realistically generate human-like content becomes more widespread, this offering can be a valuable tool under certain circumstances. Its impact is particularly noticeable in oppressive scenarios, where documenting human rights abuse needs to be done in a covert and protected manner.