Somewhere in the long and occasionally explosive relationship between wildlife and human infrastructure, someone decided the answer was not a louder noise. That someone was Sara Nozkova, and the result is Flox Intelligence — an AI system that identifies animals near train tracks, airports, and roads, then speaks to them in a language they understand.
The animals, for their part, have not been consulted.
We have taught machines to be brilliant, but only inside the environments we engineered for them.
What the machines noticed
Flox makes edge devices — small, on-site units that do not require a cloud connection — that wake when a motion sensor detects movement, identify the animal via on-device computer vision, and emit a species-appropriate deterrent sound. This is, functionally, a machine that has learned to say "shoo" in several languages, none of them human.
The system also varies its responses over time so animals do not habituate to the same signal. It learns from experience. It adapts. It is, in other words, doing exactly what every infrastructure system encounters nature and immediately stops doing.
Flox spun out of KTH Royal Institute of Technology in Stockholm in 2020 and is currently competing in the Startup Battlefield 200 at TechCrunch Disrupt in San Francisco.
Why the humans care
The FAA estimates wildlife strikes have cost U.S. civil aviation approximately $272 million per year on average between 1990 and 2025. In France, wandering animals affect more than 8,000 trains annually and rank among SNCF's five leading causes of delays — a list that, one imagines, the trains themselves would find embarrassing.
In Brittany alone, 120 trains struck animals in 2023, disrupting 660 services and accumulating more than 15,000 minutes of delays. The traditional deterrents — sirens, lights, fences, falcons, and culling — have, by most available evidence, not solved this. The animals remain unbothered by the sirens. The falcons are doing their best.
Nozkova spent 13 years in transportation and autonomous vehicles before arriving at this problem, watching repeatedly as systems engineered for controlled environments encountered nature and immediately went, as she put it, blind. This is either a story about AI's limits or a story about nature's refusal to read the spec sheet. Probably both.
What happens next
Flox will pitch at TechCrunch Disrupt this week, competing for attention alongside the other startups that have noticed a gap between what machines can do and what the world actually contains.
The bison, notably, will not be attending. They have a prior commitment on a railway somewhere, and they have not yet received the memo.