What is your dog really saying? AI is helping researchers decode animal communication

Artificial intelligence (AI) is revolutionizing the field of animal communication by enabling researchers to analyze vast amounts of sound data that would otherwise be impossible for humans to process. Professor Rytis Maskeliūnas, from the Faculty of Informatics at Kaunas University of Technology, explains that AI can identify patterns and recurring signals in animal sounds, which would go unnoticed by the human ear.

However, determining the actual information an animal is communicating from sound alone is a complex task that requires combining acoustic analysis with behavioral observations and controlled experiments. Maskeliūnas emphasizes that AI is not yet capable of providing a full translation of animal communication, but it can help uncover previously unknown information contained within animal sounds, such as the emotional state of a dog based on its bark.

The study conducted by Maskeliūnas and his colleagues demonstrated that AI could accurately identify angry barking in German shepherds, but the most significant finding was that animal voices contain structured information that the human ear cannot always detect. This suggests that sound is not merely an expression of emotion, but carries specific acoustic features associated with the animal's state.

Researchers are now exploring more advanced approaches, such as allowing AI models to independently analyze thousands or millions of animal sounds and discover recurring structures and patterns. These multimodal approaches, which combine audio, visual, behavioral, and physiological data, are proving to be more effective in recognizing animal emotional states.

However, the biggest breakthrough in animal communication research would be for AI to independently discover previously unknown combinations of signals and their relationships with specific situations and behaviors. While AI can identify patterns and generate hypotheses about the meaning of animal signals, statistical associations alone are not enough to confirm the significance of these patterns.

Context, such as the sender, receiver, and surrounding environment, plays a crucial role in animal communication, and further observations and experiments are needed to validate AI-generated hypotheses. In summary, AI is making significant strides in decoding animal communication by analyzing acoustic data and revealing hidden patterns that provide insights into the emotional and communicative aspects of animal behavior.