The US Navy endeavored to develop a covert code using whale songs, dolphins, sea lions, and seals during the Cold War period. This initiative involved broadcasting these marine mammal noises from submarines and decoding them using a computer called the Combo Signal Recognizer (CSR) at the receiving end. However, challenges arose with the authenticity of the code due to background noise in the recordings and the difficulty of avoiding patterns that would be easily detected by sonar operators
Despite the seemingly straightforward concept, technical limitations and complexities led to the project's failure. Attempts to synthesize whale sounds for communication purposes continued post-Cold War, with advancements in deep learning neural networks enabling the creation of synthetic data for various applications like underwater communication and sonar construction. The idea of using synthetic whale songs as a covert communication method faced criticisms, highlighting the potential negative impact on marine life and the inherent challenges in translating human communication into natural signals for animals like whales. The comparison of communication layers between humans and whales underscored the significant differences in information processing, making it unlikely for human-designed codes to effectively mimic natural signals comprehensible to whales. https://www.schneier.com/blog/archives/2024/04/whale-song-code.html