Nguyen Hoang Gia Bao

Abstract

The rapid spread of artificial intelligence (AI), especially generative AI, is changing how tasks are carried out, how workers learn and how work performance is judged. However, access to AI does not automatically lead to higher productivity. The outcome depends on whether workers can understand the limits of AI, give clear instructions, integrate AI into their workflows, verify outputs, protect data and keep meaningful human control. This study develops an integrative framework that explains how workplace AI competence affects the work performance of young workers in Vietnam. The framework draws on human capital theory and the task-based approach, task-technology fit theory and the automation-augmentation paradox. Methodologically, the study combines a structured integrative literature review, an analysis of legal and policy documents, and official secondary labour market data for the period from 2021 to the first half of 2026. International experimental evidence shows that generative AI can reduce task completion time, improve output quality and raise productivity in many knowledge tasks, but these gains weaken or even reverse when tasks lie outside the capability frontier of the technology or when users lack verification skills. In Vietnam, the youth (15-24) unemployment rate increased from 7.63% in 2023 to 8.64% in 2025 and 8.77% in the first half of 2026, while the share of workers holding formal training qualifications continued to rise, which signals growing skill-transition pressure. At the same time, Vietnam’s legal framework for AI expanded very quickly in 2025-2026. We propose a six-dimension workplace AI competence framework, six testable propositions and a policy package supported by a three-tier system of key performance indicators (KPIs). The paper offers a conceptual model that can be tested in future empirical studies of young Vietnamese workers.

Keywords: AI competence; AI literacy; work performance; young workers; human-AI collaboration; task-technology fit; Vietnam.

1. Introduction

Artificial intelligence (AI) has moved from being a specialised technology to a general-purpose tool that can take part directly in writing, data analysis, programming, customer service, design, research and decision support. Earlier waves of computerisation mainly replaced routine manual and clerical tasks while complementing non-routine analytical work (Autor et al., 2003). Generative AI is different because it also affects language-based and knowledge-intensive tasks that used to be seen as the comparative advantage of educated workers (Eloundou et al., 2024). The early empirical evidence is striking. In a randomised experiment with 453 college-educated professionals, Noy and Zhang (2023) found that access to ChatGPT reduced the time needed to complete mid-level professional writing tasks by about 40% and raised output quality by about 18%. In a large field setting, Brynjolfsson et al. (2025) followed 5,172 customer-support agents and found that an AI assistant increased productivity, measured as issues resolved per hour, by 15% on average, with the largest gains among less experienced and lower-skilled workers. These results suggest that the value of AI depends less on whether a firm “has AI” and more on whether its workers can turn AI into a measurable, complementary capability.

However, it would be a mistake to treat “using AI” as the same thing as “performing better”. AI can produce answers that look convincing but are factually wrong; it can increase speed while reducing accuracy; and it can weaken professional judgement if users hand over too much of their thinking to the system. Dell’Acqua et al. (2023) describe this situation as a “jagged technological frontier”: the same AI model can perform very well on some tasks but fail on other tasks that look similar. In their field experiment with 758 consultants at Boston Consulting Group, consultants who used AI on tasks inside the frontier completed 12.2% more tasks, worked 25.1% faster and produced results of more than 40% higher quality than a control group. On a task deliberately chosen to lie outside the frontier, however, AI users were 19 percentage points less likely to reach a correct solution. For this reason, the key variable to study is not how often people use AI but their workplace AI competence, that is, the combination of knowledge, operational skills, evaluation ability, workflow redesign ability and responsible conduct that allows a worker to achieve better results without sacrificing quality, legal compliance or human control.

This issue is especially important for young workers. People at the beginning of their careers are usually highly exposed to digital technology, learn new tools quickly and are less attached to old routines. At the same time, they have less domain experience, tacit knowledge and situational judgement than senior colleagues. Young workers may therefore gain the most from AI, but they are also more exposed to risks such as over-reliance, undetected errors and the loss of the entry-level tasks through which professional skills are normally built. The OECD (2023) reports that skill shortages are one of the main barriers to AI adoption in firms and that training is needed not only for low-skilled workers but also for high-skilled workers and managers. At the global level, Gmyrek et al. (2025) estimate that about one in four workers is employed in an occupation with some exposure to generative AI, yet they conclude that the transformation of jobs, rather than their full replacement, is the most likely outcome because most occupations still require human input and oversight. Employers also expect fast skill change: according to the World Economic Forum (2025), 39% of workers’ core skills are expected to change by 2030, and AI and big data are the fastest-growing skills.

In Vietnam, the period 2021-2026 saw very fast institutional change. The first National Strategy on AI was issued in 2021 (Prime Minister of Vietnam [PM], 2021), and Resolution No. 57-NQ/TW of December 2024 made science, technology, innovation and digital transformation the key drivers of national development (Politburo of the Communist Party of Vietnam [Politburo], 2024). In 2025-2026, a new layer of institutions was added, including special pilot mechanisms for innovation (Government of Vietnam [GoV], 2025; National Assembly of Vietnam [NA], 2025e), the Law on Science, Technology and Innovation (NA, 2025b), a new Law on Employment (NA, 2025d), the Law on Personal Data Protection (NA, 2025a) and, most importantly, the Law on Artificial Intelligence, in force since 1 March 2026 (NA, 2025c), with Decree No. 142/2026/ND-CP guiding its implementation (GoV, 2026). In August 2026, the Prime Minister also approved a National Programme for AI Human Resource Development (PM, 2026b) and a new National Strategy on AI to 2030, which replaced the 2021 strategy (PM, 2026c). As a result, the AI competence of workers is no longer only a voluntary technical skill; it is increasingly tied to obligations of risk management, transparency, accountability and human oversight.

Article details here