Artificial intelligence (AI) has not only changed the way people work, but also subtly influenced their way of thinking. Nowadays, AI can almost independently complete complex tasks such as contract drafting, disease diagnosis, tumor detection, artistic creation, and even code writing. However, hidden worries are gradually emerging behind the prosperity. The loss of job skills, degradation of social interaction skills, and weakened ability to analyze, understand, and judge key issues caused by AI have attracted the attention of many scientists. The latest research by Anastasia Berg, assistant professor of philosophy at the University of California, Irvine, and others shows that employees who overly rely on AI are losing their core skills at an alarming rate. The professor warns that companies competing to introduce AI to enhance productivity may be leading to the bitter fruit of hollowing out human core expertise. AI is stealing core skills and bringing convenience and efficiency to humans in many fields, even surpassing humans in tasks such as reading X-rays and summarizing complex documents. However, the latest research from authoritative publishing institutions such as Oxford University Press and Springer shows that although AI can improve learning speed, it often comes at the cost of sacrificing deep thinking, critical thinking, creativity, and long-term skill development. Aniket Quetel, a professor at Carnegie Mellon University's Institute of Human Computer Interaction, pointed out that seemingly beneficial AI advances in the short term, especially improvements in work efficiency, may lead to more profound problems, namely the hollowing out of core expertise in various industries. Professor Berg specifically pointed out that junior employees are most susceptible to the impact of skill loss. Due to their reliance on AI from the beginning, they were unable to establish the fundamental knowledge required to understand AI behavior, let alone verify or correct AI output. Professionals are also not immune. A study published in The Lancet Gastroenterology and Hepatology in October this year found that doctors who are accustomed to using AI assisted colonoscopy performed worse without AI support: the detection rate of precancerous lesions decreased from 28.4% to 22.4%. Similar phenomena are spreading in fields such as law, education, news, and software development. A professor at Illinois Law School in the United States has found that students who use generative AI tools are more likely to make critical mistakes. They warn that without proper checks and balances, this technology may lead to widespread skill loss, especially among young and inexperienced lawyers. Berg further stated that AI dependency has surpassed the realm of the workplace. Adults' dependence on chatbots for emotional support and daily decision-making is weakening their ability to make independent judgments and cognitive skills required for independent living. The reason for the loss is thought-provoking. A study by Microsoft scientist and Carnegie Mellon University doctoral student Hank Lee in 2025 revealed that generative AI makes task execution seem easy, but it leads people to give up their professional knowledge of problem-solving and focus on functional tasks such as collecting and integrating responses. At the same time, their confidence in using AI has increased, and this high level of confidence may make them believe that reducing their level of effort is reasonable. Berg strongly agrees with this. She pointed out that AI not only automates tasks, but also automates the entire process of developing skills for people. Once employees become dependent, they lose the opportunity to exercise their reasoning, problem-solving, and decision-making abilities. She emphasized that if companies continue to introduce AI into all workflows under the guise of efficiency, they may cultivate a generation of employees who are "superficially efficient" but at a loss for AI assistance. In other words, AI may not only fail to improve the level of labor, but also quietly erode it. Janet Frances Ravner, a researcher at the Department of Management at Aarhus University in Denmark, also warns that if left unchecked, skill loss will erode individuals' professional abilities and organizations' core competitiveness. However, Jacob Shesen, the director of the school's Hybrid Intelligence Center, has reservations about the claim that humans are on the brink of "AI reducing human efficiency," but he is also concerned about the decline in skills and its consequences, which often only become apparent after the fact. In the face of challenges in the field of hybrid intelligence, OpenAI Kokotailo team member Thomas Sherson proposed the solution of "hybrid intelligence", which is to create a better future by learning new skills and improving existing ones. Shelsen et al. suggest that organizations must establish measurement indicators to simultaneously track technology applications and human capacity building. This includes assessing factors such as employees' AI self-efficacy and psychological safety, with the goal of placing humans upstream in the value chain. Associate Professor Matt Bean from the University of California, Santa Barbara, stated that organizations must design AI technologies around skill development. Microsoft Senior Researcher Adwit Salka believes that organizations must help employees adapt to change, and training programs should focus on developing new critical thinking skills for the use of AI. Kevin Croston, a professor of information science at Syracuse University, emphasizes that certain decisions, especially those involving the interests of others, must be made by humans. Society should also adhere to the bottom line of core skills that cannot be compromised, including the ability to verify calculations, write and analyze information clearly. In addition, humans must maintain core competencies in important areas and take responsibility for their actions. Experts unanimously believe that the key to success in the AI era lies in understanding its dual nature of increasing productivity but potentially reducing skills, and wisely seeking a way for humans and AI to coexist harmoniously: actively unleashing AI's potential while safeguarding human basic skills. (New Society)