AI and drones are decoding whale calls. What they're finding could change how we think about animal culture.

Key points

• Researchers used hydrophones, underwater microphones, to record beluga whale calls in the Canadian Arctic from 2015 onward.
• Belugas produce dozens of distinct sounds, which scientists believe help them stay connected in dark, murky Arctic water.
• AI tools now help researchers sort and analyse thousands of hours of whale audio far faster than any human team could.
• Humpback and sperm whales also show signs of learned vocal traditions passed between individuals.
• Scientists say understanding animal communication could reshape what we know about culture and language outside of humans.

In 2015, whale researcher Valeria Vergara pitched a small tent at the edge of an icy bay on Somerset Island, deep in the Canadian Arctic. Fog was so thick she could barely see her hands. She dropped a hydrophone, an underwater microphone, into the freezing water and waited.

Up to 1,000 beluga whales, many travelling with newborn calves, filled the surrounding bay with near-constant noise: clicks, chirps, whistles, squeals. Dozens of different sounds, all from animals whose survival may depend on staying in contact through dark, cloudy Arctic water.

Why does any of this matter to the rest of us?

This is where AI enters the picture, turning a story about remote Arctic fieldwork into something with broader stakes.

Manually listening to thousands of hours of underwater recordings is slow and prone to human error. Researchers now feed those recordings into machine learning systems, a type of AI that finds patterns in large data sets, to sort and label calls far faster than any human team could manage. What once took years can now take weeks.

The Guardian first reported on this wave of AI-assisted wildlife research, which is uncovering evidence that whales don't just communicate. They may actually teach each other.

Humpback whales appear to pick up new songs from neighbouring groups, spreading specific call patterns across ocean populations the way a catchy tune spreads through a crowd. Sperm whales use distinctive rhythmic clicking patterns, known as codas, that differ between social groups called clans, and young sperm whales seem to learn their clan's coda rather than being born knowing it.

Learned behaviour passed through a social group is one of the defining features of culture. If whales do it, the line between human culture and animal behaviour gets a lot blurrier.

What does this mean for whale conservation?

Decoding communication could directly help protect species at risk. Beluga populations face pressure from climate change and shipping noise. If scientists can identify calls linked to feeding or mother-calf bonding, wildlife managers gain a practical tool: acoustic monitoring that can detect distress in real time, without expensive boat surveys.

Drones add another layer, letting researchers film whale behaviour above water at the same moment hydrophones capture what's happening below, matching sounds to actions in a way that was nearly impossible a decade ago. We reported on 25 August 2026 how NVIDIA's new Jetson Orin Nano 2 chip aims to put this kind of AI processing directly inside physical machines like drones, which hints at how quickly the hardware behind field research like this is moving.

Common questions

Do whales actually have a language?

Not in the way humans do. Whale communication is complex and learned, but scientists haven't yet established that it carries the grammar or open-ended meaning that defines human language. It's far richer than simple instinct, though.

How does the AI actually help researchers?

It acts like a fast librarian. The system listens to recordings, groups similar sounds together, and flags patterns, freeing biologists to focus on what those patterns mean rather than spending years cataloguing raw audio.

Can this research really help save whales?

Potentially, yes. Identifying stress calls or disrupted communication patterns could give conservationists an early warning system for threats like shipping traffic or habitat change, making protection efforts faster and more targeted.

Here's my honest read on this: the AI isn't replacing the scientist in the tent at the water's edge. It's giving her a far better chance of understanding what she hears, and that's the part worth watching.