Grey whales: an AI thermal camera to avoid ship strikes

In San Francisco Bay, a new technology has just come into service to protect grey whales from ship strikes. Called WhaleSpotter, the system combines thermal cameras and artificial intelligence to spot cetaceans by day and by night, up to seven kilometres away, and to warn vessels in time. A technological response to an ecological crisis that is hitting the species hard.

Starving grey whales, struck by ships

In 2025, twenty-one grey whales were found dead in San Francisco Bay and the surrounding waters; about 40% of them bore the marks of ship strikes. Along the whole west coast of the United States, hundreds of individuals have stranded, many of them severely malnourished.

The demographic picture is just as alarming: the Eastern North Pacific grey whale population fell from about 20,500 individuals in 2018 to only 14,500 in 2023, a drop of nearly a third in five years. A decline that can be set in context with our population data on cetaceans.

Climate change at the root of the problem

Why do these whales venture into such a busy bay? The answer goes back to the Arctic. As it melts, the sea ice deprives the algae that live beneath it of light. Those algae feed the amphipods, small crustaceans of the sediment on which the grey whale gorges during the summer. With that resource collapsing, starving animals change their route and push into San Francisco Bay in search of food, straight into the shipping traffic.

This mechanism shows concretely how warming is upsetting food chains and the migration routes of cetaceans, a phenomenon we document in our database on the impacts of climate change.

WhaleSpotter: a thermal camera trained by AI

Developed over nearly fifteen years by the Woods Hole Oceanographic Institution (WHOI), the technology rests on thermal imaging. A whale at the surface is about 2 °C warmer than the water around it: its blow and its back give away a heat signature that an infrared camera can pick up continuously, including at night.

The system was trained on hundreds of thousands of thermal images to tell a whale from a boat, a wave or a bird. When a detection is confirmed, it alerts vessels so that they slow down or change course. “Shipping is not going to go away”, sums up Daniel Zitterbart, a physicist at WHOI and scientific lead on WhaleSpotter: the task is therefore to make boats and whales coexist. This approach is part of the rise of cetacean tracking technologies, from drones to artificial intelligence.

Concrete deployments in San Francisco Bay

A first camera was installed in May 2026 on a radio tower on Angel Island; a second is fitted to a passenger ferry that crosses the bay. Other sites are under consideration, such as the Golden Gate Bridge or Alcatraz, to cover the riskiest crossing points.

The project brings together a broad coalition: the Benioff Ocean Science Laboratory at the University of California, Santa Barbara, led by the ecologist Douglas McCauley, the United States Coast Guard, whale monitoring specialists and local ferry companies. Ship strikes are indeed among the leading causes of death in large cetaceans.

What to take away from this innovation

WhaleSpotter will not settle the fate of grey whales on its own: the underlying cause, the warming of the Arctic, reaches far beyond San Francisco Bay. But by reducing collisions, the technology may buy valuable time for a population in steep decline.

Above all it illustrates an underlying trend: the growing use of artificial intelligence and remote sensing in marine conservation. A subject to follow closely, since the risk of collision also threatens the large cetaceans of the Mediterranean, where fin whales and sperm whales are found.

Source: after “AI and other tech could help save imperiled gray whales”, Science News, 2026.