New Research Shows Pigeons Solve Problems the Same Way Artificial Intelligence Does
Michelle Milliken
Pigeons have been successfully trained by humans to do a variety of things, from carrying messages in war time to racing, A new study shines some light onto how these teachable animals think.
Research recently published in iScience took a look at how pigeons solve visual categorization tasks, something at which they’re particularly gifted. The goal was to see if they’re able to master these tasks through complex processes. Throughout their experiments, the team learned that the birds solve such tasks in an interesting manner, though: in a relatively simple way that mirrors how artificial intelligence learns.
Brandon Turner, lead author and professor of psychology at The Ohio State University, says, “We found really strong evidence that the mechanisms guiding pigeon learning are remarkably similar to the same principles that guide modern machine learning and AI techniques.
“Our findings suggest that in the pigeon, nature may have found a way to make an incredibly efficient learner that has no ability to generalize or extrapolate like humans would.”
The study involved showing pigeons items that included various angles, widths, and types of rings to see if they could categorize them. The birds pecked a button on the left or right to show which category they were choosing on each task. When they were right, they’d get a treat.
The team found that with trial and error, the pigeons were able to gradually increase their success on these experiments. Their accuracy on the easiest tasks increased from 55 to 95%, and even on the tougher ones, their scores rose from 55 to 68%.
The researchers say this gradual success might have been due to associative learning, which involves making a connection between two things, like when your dog learns that sitting and getting a treat go together. This is a pretty easy thing to teach, though, not something complex, which may show the categorization doesn’t require a higher thought process.
When the researchers then observed how AI learned the same task, they found it did so in the same trial and error and associative learning method as the pigeons, meaning that a basic thought process like associative learning may help with the mastering of more complex categorization tasks.
How does this relate to how humans would tackle such tasks? The researchers say we look for rules to make things easier. Turner explains that there weren’t any rules that could do that, though, so we’d likely give up on it. For pigeons, they just kept using “brute force” to power through trial and error and associative learning, much like AI.
The team says their findings indicate that humans should feel a little less intellectually superior to pigeons.
Turner explains, “We celebrate how smart we are that we designed artificial intelligence, at the same time we disparage pigeons as dim-witted animals.
“But the learning principles that guide the behaviors of these AI machines are pretty similar to what pigeons use.”
So the next time you see a pigeon, don’t assume they don’t have the smarts to get things figured out.
