How AI Is Helping Scientists Protect Wildlife Through Sound

• Researchers are using AI to analyze massive datasets from autonomous sound-recording devices to monitor global biodiversity more efficiently.
• The EU-funded BioacAI project, led by Professor Dan Stowell, aims to bridge the gap between acoustic data collection and actionable ecological insights by 2027.
• The initiative includes a doctoral network to train experts in acoustics, AI, zoology, and ecology.

The Data Challenge

• Passive acoustic monitoring generates hundreds of terabytes of data annually, which is impossible to process manually.
• Existing traditional field surveys are too labor-intensive, costly, and limited in scale to combat accelerating biodiversity loss effectively.
• BioacAI is developing smarter devices that can run recognition algorithms on-site to reduce power usage and data processing backlogs.

Focusing on Bats

• Bats are difficult to track due to their nocturnal, elusive nature; passive acoustic monitoring is the primary method for studying their populations.
• Distinguishing between bat species via echolocation is challenging because calls adapt to environments.
• The team is investigating bat "social calls"—which are often more species-specific—to improve automated classification accuracy.

AI Innovation and Discovery

• Researchers use "deep embeddings" to map similar animal sounds together, enabling the system to identify known species and flag unusual, unclassified sounds for human investigation.
• This approach aims to move beyond monitoring common species toward detecting biodiversity trends, habitat shifts, and new ecological hotspots.
• The technology supports the EU Biodiversity Strategy for 2030 by providing policymakers with clearer data on ecosystem changes.