On a ferry gliding through San Francisco Bay, a thermal camera mounted on deck watches for something invisible to the naked eye: the heat signature of a whale's breath breaking the surface. When it detects one, artificial intelligence springs into action, alerting the ship's crew within minutes through the Whale Safe website. It's a simple chain of human and machine intelligence designed to solve a collision crisis that emerged only recently—but with urgent consequences.
Starting in 2018, gray whales began making an unexpected stop in one of America's busiest ports. These marine giants, which undertake one of the longest migrations of any mammal, traveling roughly 19,000 kilometers from feeding grounds in Alaska to breeding grounds in Mexico and back, had never regularly visited San Francisco Bay before. Scientists believe climate change has made their Arctic feeding grounds less productive, leaving the whales hungrier as they head south to breed. The bay offers a convenient meal during their journey—but it also places them in the crosshairs of modern shipping. In 2025 alone, more than 20 gray whales were killed by ship collisions in the bay.
Enter the Whale Safe project, a collaboration between the Benioff Ocean Science Lab, WhaleSpotter, and the Marine Mammal Center. Researchers have deployed thermal cameras that detect the heat signature of whale spouts and bodies when they surface, bypassing the need for human observers to be watching at every moment. "Next a trained human confirms the detection and will help classify the species when possible," explains Rachel Rhodes, a project scientist with the Benioff Ocean Science Lab. The confirmed sightings are posted publicly on the Whale Safe website, accessible to mariners, the Vessel Traffic Service, and ferry operators throughout the Bay Area. They can then adjust their speed or course to avoid the whales.
Currently, the project operates with two thermal cameras mounted at strategic points. One sits on a Coast Guard tower on Angel Island in the middle of the bay, while the other is installed on a ferry that passes through a known hotspot for gray whales. The advantage over human spotters is clear: thermal cameras with AI detection systems work 24 hours a day, never tiring, never looking away. The researchers envision scaling this network significantly, with cameras eventually mounted on bridges, ferries, and other infrastructure throughout the Bay Area to provide comprehensive coverage of the entire bay.
Before that expansion happens, Rhodes notes the team will spend the next two seasons consulting closely with "Bay Area stakeholders to ensure the data from these systems is being shared in a way that meets their needs and is giving them all the insight they need to avoid collisions with whales." This careful, collaborative approach reflects the broader complexity of shipping in a major port—balancing commerce with conservation requires more than technology alone.
The project has already captured attention at the state level. California State Representative Sam Liccardo introduced legislation to create a "whale desk" dedicated to protecting the whales while helping mariners avoid what he described as "costly, harrowing collisions." As climate change continues to reshape migration patterns and push hungry whales into new waters, this pilot project in San Francisco Bay offers a model for how technology, science, and human communication can work together to protect both marine life and maritime operations.