The test for any breakthrough technology is often where you least expect it, but once it “conquers” that application, even more possibilities may emerge.
We all know that artificial intelligence (AI) is touching every aspect of our lives. From when we turn on our smartphone to searching the web to making online purchases, AI is helping us make better decisions, and ideally, make better use of our time.
And yet, one area that AI is having a significant impact on is birding.
Birding?
Birding is the recreational activity of looking for and identifying birds. Birders (as they call themselves) get up before daybreak to look for birds, just when the birds are most active. Some birders exhibit an obsession-like thirst to add to their life list of birds. In North America, there are more than 800 species of birds. Getting a glimpse of every such bird species is a daunting task. Yet birders will spend a great deal of time (and money) searching for species that have eluded them.
Some birders have what has been called a “Big Year,” when they strive to see as many species as possible within a calendar year.
Then there are fall (which is now in full throttle) and spring migrations. Some birds spend most of their life flying between South America where they winter and North America where they breed in the summer. The water around Pittsburgh makes it an attractive birding area during migration. These periods represent fleeting opportunities for birders to see birds which pass through their area for short periods of time.
So what does this have to do with AI?
Every bird has a unique song (or vocalization). Birders like to see every species, but they are more likely to hear a species before they see it. The challenge is being able to identify the sound attached to each bird species.
That is where the Cornell Ornithology Laboratory comes in. The lab created Merlin, an AI app that can be downloaded at no cost on your smartphone. When when turned on it will listen for bird songs and give you its best guess of the bird species. It uses machine learning to compare bird song spectrograms to the bird song that you are hearing. By comparing such sound footprints, Merlin gives you an indication of what the bird you are hearing may be.
My wife is a birder. I would classify her as mildly obsessed. We spent a good deal of our honeymoon searching for special bird species on the West Coast.
Recently I joined her on one of her birding excursions and she showed me Merlin. By letting Merlin suggest what birds we heard, she was able to match Merlin flawlessly.
My attention on our birding outings is focused entirely on Merlin, and how well my wife can name a bird. This is akin to the popular game show “Name that Tune.” with bird songs rather than popular songs the target of identification.
As a computer scientist, Merlin has caught my attention. I will never be sufficiently enthralled with birding to get up at 4AM to see what species are around. I will however turn Merlin on anytime I am outdoors, because to some, Merlin seems like magic. To a computer scientist like myself, Merlin demonstrates how AI can impact an activity like birding, enhancing the power to identify bird species. Merlin makes a simple hobby like birding into an AI experience.
So yes, thankfully, even AI is for the birds, and birders.
