Scientists Use Artificial Intelligence (AI) to Decode Sperm Whale Communication

Key facts

• Researchers using Artificial Intelligence (AI) have identified a possible phonetic alphabet within the complex communication system of sperm whales.
• The Cetacean Translation Initiative (Project CETI) analysed recordings of approximately 400 sperm whales in the Eastern Caribbean captured between 2005 and 2018.
• Project CETI mapped 156 distinct codas, identifying structural adjustments known as rubato (manipulation of tempo and rhythm) and ornamentation (appending an extra click to a sequence).
• A study published in the journal Proceedings of the Royal Society B: Biological Sciences revealed that Mediterranean sperm whales have developed distinct regional dialects.
• The endangered Mediterranean sperm whale population, numbering between 250 and 2,500 mature individuals, descended from a group that entered the sea through the Strait of Gibraltar roughly 20,000 years ago.
• Whales in the western Mediterranean near Spain's Balearic Islands maintain a traditional 3+1 coda, while those in the eastern Mediterranean near the Hellenic Trench off Crete use a sped-up version of the same sequence.
• Data indicates that eastern Mediterranean sperm whales possess the ability to code-switch to the western dialect depending on context or contact.