Pigeons puzzle experts with AI-matching intelligence
While the world is impressed with AI capabilities, scientists tried pigeon intelligence only to conclude that they can learn similarly to AI.
Researchers at the University of Iowa recently discovered that pigeons’ learning abilities match those of artificial intelligence (AI) as they achieve almost 70% accuracy in association learning tests.
The research findings have shed light on the learning abilities of birds, despite the popular mockery of feral pigeons. The birds can solve “diabolically difficult” tasks, as described by the lead scientist in the research – Dr. Ed Wasserman.
“The pigeons are like AI masters,” said the scientists, explaining that both pigeons and AI are capable of associative learning.
“They’re using a biological algorithm, the one that nature has given them, whereas the computer is using an artificial algorithm that humans gave them,” said Wasserman, who has spent the past five decades researching pigeons.
Test pigeons showed impressive results
The experiment involved a group of four pigeons subjected to a series of visual stimuli. The test pigeons were required to peck a button on either their right or left side based on the category to which each stimulus belonged. Correct responses were rewarded with food, while incorrect responses received no reward.
According to Dr. Wasserman, the test was "diabolically difficult." The stimuli differed in characteristics such as line width, angle, and concentric or sectioned rings, making it impossible to sort them into categories using rules or logic. Dr. Wasserman explained that the pigeons had to memorize the individual stimuli in order to complete the task.
Despite the initial difficulties, the pigeons were eventually able to achieve an average accuracy rate of 68% after many hundreds of tests and the incentive of a food reward.
The results, according to Wasserman, demonstrate the pigeons' exceptional skill in "low-level" associative learning. “They’re workaholics. They didn’t quit. And they did learn,” says Dr. Wasserman.
He believes that if a similar test were administered to humans, the results would not be nearly as impressive, with many participants likely quitting due to the difficulty of the task.
Animals use the same learning algorithm as AI
The researchers in the experiment investigated two types of learning – declarative learning and associative learning. Declarative learning is a more advanced skill primarily associated with humans and involves reasoning with groups of rules or tactics.
Associative learning is a more simple technique that focuses on recognizing patterns and forming connections. Many species of animals widely use associative ways to learn.
The researchers suggest that the same associative learning type is employed by computers and artificial intelligence, which can be taught to recognize patterns and objects.
“You hear all the time about the wonders of AI, all the amazing things that it can do,” claimed Wasserman, “It can beat the pants off people playing chess, or at any video game, for that matter. It can beat us at all kinds of things. How does it do it? Is it smart? No, it’s using the same system or an equivalent system to what the pigeon is using here.”
Scientists discover capabilities of pigeon
In recent years, researchers have uncovered the remarkable cognitive abilities of pigeons. In 2015, a team of researchers at the University of California trained pigeons to detect breast cancer with surprising accuracy.
The birds were taught to distinguish between benign and malignant tumors and identify the difference between the two reliably.
The following year, New Zealand researchers conducted another experiment revealing that pigeons can accurately determine the difference between four-letter words and four letters in random order (non-words).
More from Cybernews:
SpaceX’s Starlink is a fair wartime target
San Diego healthcare provider admits data breach
Wikipedia shutdown earns Pakistan human rights slap
Man sentenced for targeting Optus breach victims
New Netflix account sharing rules explained
Subscribe to our newsletter
Your email address will not be published. Required fields are marked