The other day I saw this story about drones that work with AI to spot sharks in the water near swimmers. The article describes the technology as a win-win for people and wildlife, and I hope this is true. Certainly the technology that powers this must have some amazing applications in the protection of wildlife in all sorts of places.
The article says that the AI can identify sharks with 90% accuracy, compared with 12% for humans (though perhaps that’s because we’re likely to be driven by terror and assume that everything is a shark).
I find it really interesting that the technology that drives this is becoming so accessible. You can test out the image analysis AI at the Microsoft Cognitive Services site (it was a full 5 years out when it guessed my age, but the photo I used wasn’t the best). And you can licence Microsoft’s AI to use in your own applications. There are several companies doing just this to run services for schools that help to identify individual students in photos, or to identify people who don’t belong in the school.
Google and Apple obviously use very similar technology in their Photos applications. I guess the other interesting thing about this technology is the amount of power this gives the companies that control it; certainly more than many states.