Latest news with #LawofEffect
Yahoo
10-07-2025
- Science
- Yahoo
Letters to the Editor: Some AI models actually are learning how to solve problems like humans do
To the editor: Guest contributor Iddo Gefen not only laments the analogy between the human brain and artificial intelligence, but he also suggests that human minds don't learn or recall like an AI ('The human brain doesn't learn, think or recall like an AI. Embrace the difference,' July 9). In fact, Gefen gets it backwards: Recent large language reasoning models of AI learn, recall and solve problems much in the way humans do. Humans (and other animals) learn most of our important behaviors through reinforcement, the foundational law of learning first suggested by psychologist Edward Thorndike's Law of Effect and later experimentally investigated and confirmed by the behavioral psychologist B. F. Skinner and his colleagues. In addition, verbal humans frequently solve problems by talking (and imagining) to ourselves until we find a solution. Reinforcement plays a critical role in learning and engaging in problem-solving behavior. Reasoning LL models of AI that are also programmed to learn through reinforcement mimic such problem-solving behaviors in humans and are now among the most powerful AI machines. Unfortunately, along the way, Gefen ventures beyond his expertise as a neuroscientist and mischaracterizes 'behaviorist psychology.' If he looked deeper, however, he would find that behavioral psychologists have discovered laws of behavior that have revolutionized the treatment of a wide range of behavior disorders, as well as revolutionizing education. They now provide the foundation for a whole new generation of AI machines that think and learn like humans. Henry D. Schlinger Jr., GlendaleThis writer is a professor of psychology at Cal State Los Angeles. This story originally appeared in Los Angeles Times.
Los Angeles Times
10-07-2025
- Science
- Los Angeles Times
Letters to the Editor: Some AI models actually are learning how to solve problems like humans do
To the editor: Guest contributor Iddo Gefen not only laments the analogy between the human brain and artificial intelligence, but he also suggests that human minds don't learn or recall like an AI ('The human brain doesn't learn, think or recall like an AI. Embrace the difference,' July 9). In fact, Gefen gets it backwards: Recent large language reasoning models of AI learn, recall and solve problems much in the way humans do. Humans (and other animals) learn most of our important behaviors through reinforcement, the foundational law of learning first suggested by psychologist Edward Thorndike's Law of Effect and later experimentally investigated and confirmed by the behavioral psychologist B. F. Skinner and his colleagues. In addition, verbal humans frequently solve problems by talking (and imagining) to ourselves until we find a solution. Reinforcement plays a critical role in learning and engaging in problem-solving behavior. Reasoning LL models of AI that are also programmed to learn through reinforcement mimic such problem-solving behaviors in humans and are now among the most powerful AI machines. Unfortunately, along the way, Gefen ventures beyond his expertise as a neuroscientist and mischaracterizes 'behaviorist psychology.' If he looked deeper, however, he would find that behavioral psychologists have discovered laws of behavior that have revolutionized the treatment of a wide range of behavior disorders, as well as revolutionizing education. They now provide the foundation for a whole new generation of AI machines that think and learn like humans. Henry D. Schlinger Jr., GlendaleThis writer is a professor of psychology at Cal State Los Angeles.
