In recent years, with the continuous improvement of algorithmic computing power and the continuous accumulation of massive data, artificial intelligence is moving from a "decision-making" era to a "generative" era. Ten years ago, scholars generally believed that decision based artificial intelligence was difficult to overcome bottlenecks such as perceptual manipulation, creativity, and social intelligence in the short term, and predicted that professions with lower skill requirements would be more easily replaced. Just ten years later, generative artificial intelligence overturned this concept. Generative artificial intelligence has not only changed the nature of existing professions, but also nurtured numerous emerging professions, which not only pose new requirements for the skills of workers, but also eliminate some traditional skills. Career Task Replacement: From Decision Based Artificial Intelligence to Generative Artificial Intelligence, the design goal of Decision Based Artificial Intelligence is to simulate human decision-making processes, analyze input data and information, and then make the most likely decisions. It is usually used in scenarios that require quick and accurate decision-making, such as financial investment, medical diagnosis, etc. Generative artificial intelligence, also known as creative artificial intelligence, is an artificial intelligence that can generate new outputs similar to human creativity. Its design goal is to mimic human creativity by learning and understanding a large amount of data and information, and then generating new and innovative content, usually used in scenarios that require a lot of innovative content, such as artistic creation, news writing, etc. Compared with decision based artificial intelligence, generative artificial intelligence requires more powerful algorithm capabilities, larger data scales, and more complex tasks to execute. Generative artificial intelligence not only possesses predictive ability, but also demonstrates various abilities such as independent judgment, content creation, and social intelligence, thereby expanding the scope of intelligent applications to the field of mental labor. In the era of decision based artificial intelligence, most economists hold the concept that artificial intelligence has "conventional substitutability", that is, artificial intelligence is more likely to replace the labor force that performs routine tasks with high repeatability, while for unconventional tasks that need to face complex and ever-changing external environments, it still relies on labor. However, with the shift towards generative artificial intelligence, the scope of tasks performed by robots has begun to extend to unconventional tasks, and the responsibilities of labor have correspondingly changed. For example, generative artificial intelligence can assist art creators in collecting materials and analyzing audience preferences, thereby improving the efficiency and quality of creation. However, high-level tasks such as deep social insights, emotional expression, and artistic aesthetics still require humans to complete. Teachers can use artificial intelligence to analyze student characteristics, develop and optimize textbooks, grade assignments, and invest more time and energy in paying attention to personalized student needs and innovative teaching models. Generative artificial intelligence can replace lawyers in completing document review, case analysis, contract drafting, and basic legal consultation, but tasks such as appearing in court for defense, establishing customer relationships, and making ethical judgments still need to be personally completed by lawyers. Even though generative artificial intelligence has greatly expanded the scope of automated tasks and changed the role of humans in many professions, tasks involving deep professional knowledge, emotional understanding, artistic creation, and moral judgment still rely on humans. Artificial intelligence not only changes career tasks, but also spawns numerous emerging professions. The Occupational Classification Code of the People's Republic of China Occupational Classification Code (2022 Edition) included 97 digital occupations for the first time, and these new digital occupations accounted for 61% of the new occupations in this revision, showing the rapid growth trend of digital occupations. These emerging professions involve supervising or maintaining artificial intelligence systems, providing higher levels of creativity and decision-making. For example, artificial intelligence trainers are mainly responsible for data annotation, algorithm parameter settings, human-computer interaction design, and performance testing and tracking; Artificial intelligence engineering technicians are mainly engaged in the analysis, research, design, optimization, and management of artificial intelligence algorithms and deep learning technologies; Electronic data forensics analysts are responsible for the extraction, fixation, recovery, and analysis of electronic data; Intelligent building administrators are mainly responsible for the operation, debugging, testing, and maintenance of intelligent building systems. The changing demands for occupational skills in the era of generative artificial intelligence are triggering a transformation of career tasks, which will undoubtedly reshape the demand for employee skills in various professions. Entering the era of generative artificial intelligence, not only can high-precision and highly repetitive tasks such as data analysis, financial transaction processing, and document classification be automated, but even some basic creative or design tasks, such as video production, text editing, and program design, can be completed by artificial intelligence. Therefore, basic data input and processing skills, as well as the ability to perform simple repetitive tasks, will gradually be phased out. However, generative artificial intelligence still cannot complete tasks involving strategic planning, complex problem-solving, advanced creative design, emotional communication, complex social environments, and high professional skills knowledge. Therefore, critical thinking, innovative thinking, emotional management, interpersonal communication, and other skills required for such tasks will become increasingly important. In addition, artificial intelligence will also create new tasks and demand new skills from workers, such as digital technology applications, data interpretation, cross disciplinary cooperation, remote operations, and other abilities. In the era of generative artificial intelligence, skill demands are rapidly changing. Whether workers should specialize in a single skill or learn as many skills as possible depends on their career goals, interests, and market demands. Specializing in a single skill can help workers become experts in a certain field, drive development in that field, and provide higher value output. However, single skill workers may also find it difficult to regain their job positions due to a lack of flexibility during job transitions. Mastering multiple skills can help workers adapt to changing work environments and gain more career opportunities, but it can lead to insufficient depth of research in a certain professional field and make it difficult for them to become experts. Therefore, workers should not be limited to improving a single skill, nor should they be satisfied with learning multiple skills at a superficial level. A more appropriate approach is to develop a "T-shaped" skill structure, which involves in-depth research in a certain field while possessing basic knowledge and skills in multiple fields, enabling workers to maintain professional depth and better adapt to changes in the times. Human machine collaboration: The direction of vocational skills training in the era of generative artificial intelligence. Although generative artificial intelligence may cause technical unemployment, intellectual property and privacy security issues, it can liberate workers from repetitive, high-intensity, and high-risk physical labor as well as primary creative and design tasks. This can not only improve production efficiency and ensure human safety, but also encourage workers to invest more energy in more complex and profound tasks, thereby promoting the progress of human society. The development of artificial intelligence is unstoppable, and learning to cooperate and coexist with it is an inevitable choice for workers to adapt to the era of digital economy. The report of the 20th National Congress of the Communist Party of China proposed to improve the lifelong vocational skills training system, which pointed out the direction for solving the employment problems caused by artificial intelligence. In recent years, various levels of government departments in China have taken various measures to improve the skill level of workers, and have successively issued a series of policy documents, such as the "Opinions on Strengthening the Construction of High skilled Talents in the New Era", the "Action Plan for Enhancing Digital Literacy and Skills for the Whole People", and the "Work Plan for Enhancing Digital Skills for the Whole People". At present, China has achieved positive results in cultivating digital skills, but there are still shortcomings in training innovative creativity, emotional understanding, and interpersonal communication skills. It is necessary for workers and policy makers to work together to build a new work environment that is more suitable for human-machine cooperation. For workers, it is crucial to establish a concept of lifelong learning, which requires constant attention to industry trends, understanding of new technology development trends, and constantly updating knowledge reserves. In terms of training content, it is not only necessary to strengthen the training of digital technology application skills, understand the basic knowledge and operation methods of artificial intelligence, but also to focus on improving soft skills such as critical thinking, problem-solving ability, and interpersonal communication ability. In terms of training methods, one should be good at using artificial intelligence tools and online course resources to improve learning efficiency, leading workers to actively participate in practical projects or use simulation platforms for practical exercises, in order to enhance practical skills and application abilities. Workers can also strengthen communication with other learners and experts by joining skill learning communities or communication groups. More importantly, workers need to flexibly choose suitable learning methods and content based on their personal interests and career goals. The relevant government departments should integrate resources from all parties and build a lifelong vocational skills training platform for workers. Firstly, promote the reform of the existing education system to meet the needs of modern positions. Strengthen the cultivation of innovative, reflective, and social skills for students, correctly guide them to use artificial intelligence tools to improve learning efficiency, and standardize the use of such tools. Actively promote the transformation of teaching thinking and methods among teachers, enhance digital teaching capabilities, such as using artificial intelligence technology to construct immersive and interactive learning scenarios, and enhance the experience and fun of teaching activities. Secondly, establish a diversified training mechanism, promote cooperation among enterprises, research institutions, educational institutions, and non-governmental organizations, and jointly carry out skills training projects or hold vocational skills competitions. Finally, formulate and improve relevant laws and regulations, regulate the development and use of artificial intelligence, effectively protect the basic rights and interests of workers, and actively establish public service platforms, develop and provide training projects and courses for artificial intelligence technology, and provide diversified learning and training resources for workers. Author: Li Jincheng, Wang Linhui (Assistant Researcher at the School of Economics of Jilin University and Professor at the Center for Quantitative Economics)