Why AI Literacy Could Become a Critical Workforce Resilience Skill

Artificial intelligence is entering the workplace not only as a productivity tool, but as a source of uncertainty about careers, skills and professional identity. The study "Job Replacement Anxiety and Approach-Avoidance Job Crafting: The Roles of Regulatory Focus and AI Literacy," published in Behavioral Sciences and authored by Chuanhao Fan and Han Wang of Hohai University, examines what employees actually do when they fear that AI could diminish or replace their jobs.

The findings complicate the familiar narrative that automation anxiety simply makes workers resistant to technology. Employees facing replacement concerns can move in opposite directions: some become more defensive and narrow their roles, while others may attempt to protect their employability by learning, adapting and reshaping their work. What determines the direction appears to depend partly on employees' motivational response to the threat, and importantly, on how well they understand AI.

Based on three-wave survey data from 583 employees in China, the research suggests that AI literacy acts as a psychological and practical resource. Greater AI literacy does not eliminate fears about replacement, but it weakens the tendency for those fears to suppress growth-oriented motivation or push employees toward defensive behavior.

The Real Workplace AI Divide May Be Between Growth and Self-Protection

The study approaches job replacement anxiety as a threat to resources employees value: income, job security, occupational identity, skill relevance and future career opportunities. When people believe those resources are at risk, they may either try to acquire new resources or conserve what they already have. That distinction becomes central to understanding how workers respond to AI.

The researchers distinguish between "approach job crafting" and "avoidance job crafting." Approach crafting involves proactive adjustment: learning new skills, adopting technology, expanding responsibilities or redesigning work relationships. Avoidance crafting instead emphasizes protection: reducing difficult demands, withdrawing from uncertain tasks or narrowing the boundaries of one's role.

The important finding is that replacement anxiety does not produce one uniform behavioral response. Its overall association with approach job crafting was not statistically significant because competing forces appeared to offset each other. Anxiety weakened growth-oriented motivation, yet there was also an opposing positive direct association suggesting that some employees may respond by trying harder to remain valuable.

The avoidance pathway was much clearer. Job replacement anxiety was positively associated with avoidance job crafting, indicating that fear of AI substitution can push employees toward reducing effort in uncertain areas, protecting existing resources and restricting their roles. For managers, this matters because the behavioral cost of AI anxiety may not always appear as open resistance; it may surface as quieter withdrawal from experimentation, challenging assignments or organizational change.

Anxiety Changes Motivation Before It Changes Behavior

The study explains what happens between feeling threatened and changing one's work. The authors examine two forms of regulatory focus: promotion focus, centered on advancement and gains, and prevention focus, centered on safety, responsibility and avoiding losses. These motivational orientations help explain why the same technological threat can produce contrasting forms of adaptation.

Job replacement anxiety was negatively associated with promotion focus. Employees worried about becoming obsolete may devote more attention to whether their jobs remain secure and less attention to advancement, experimentation or future opportunities. In turn, stronger promotion focus was closely associated with proactive job crafting, meaning that reducing this growth orientation can indirectly suppress efforts to expand skills or roles.

Replacement anxiety strongly increased prevention focus. That shifts attention toward avoiding mistakes, minimizing losses and protecting security, which was associated with more avoidance-oriented job crafting. The study therefore presents AI anxiety not simply as an emotional reaction, but as a motivational process capable of altering how employees allocate effort and respond to organizational change.

At the average level of AI literacy, the indirect effect of replacement anxiety on approach crafting through reduced promotion focus was negative, while the indirect effect on avoidance crafting through greater prevention focus was positive. The overall effect on approach crafting disappeared because the study identified an opposing positive direct association, producing what the authors describe as an inconsistent mediation pattern.

This is significant for organizational strategy. Employees worried about AI may simultaneously feel less optimistic about growth and more pressure to prove their continued value. Treating all anxiety as resistance therefore risks misunderstanding the workforce response: some workers may withdraw, while others may accelerate reskilling or expand their contribution precisely because they feel threatened.

AI Literacy Does More Than Teach People How to Use Tools

The researchers define AI literacy broadly, encompassing understanding, use, evaluation and ethical awareness rather than simple familiarity with software. Employees with stronger AI literacy may be better able to judge where AI can substitute for tasks, where human judgment remains necessary and where human-AI collaboration creates new opportunities.

The empirical results indicate that AI literacy weakened both major psychological pathways. As AI literacy increased, the negative relationship between replacement anxiety and promotion focus became smaller, while the positive relationship between anxiety and prevention focus also weakened. The corresponding indirect effects on proactive and avoidance-oriented job crafting diminished as well.

However, AI literacy was a buffer, not a cure. Even among more AI-literate employees, replacement anxiety continued to activate prevention-oriented responses. That distinction matters because training alone cannot compensate for genuine insecurity created by restructuring, unclear workforce strategies or poorly communicated automation plans.

For employers, the implication is that AI capability-building should extend far beyond short courses on prompting or tool operation. The study recommends training that includes basic AI knowledge, job-specific applications, evaluation of outputs, human-AI collaboration and ethical risks. It also points toward cross-functional learning and practical opportunities for employees to experiment with AI in real business settings.

As governments and companies invest in AI skills, literacy programs may serve not only productivity objectives but workforce resilience. In emerging economies and sectors experiencing rapid digital transformation, workers who understand the technology may be better positioned to interpret automation realistically rather than seeing it exclusively as an uncontrollable threat.

The AI Transition Is a Management Challenge, Not Just a Technology Rollout

The study suggests that organizations should treat job security communication as part of AI implementation rather than as a separate human-resources issue. Employees need clarity about which tasks are changing, what new capabilities will be rewarded and where human judgment, creativity or relational skills remain important. Reducing uncertainty may be as important as introducing the technology itself.

Management practices also shape whether workers respond defensively or constructively. The authors recommend greater autonomy, employee participation in AI pilots, room for experimentation and positive feedback, while employees displaying withdrawal may require additional training time, mentoring or task redesign. These recommendations are presented cautiously because the study is observational rather than experimental.

The three-wave design strengthens the analysis by separating measurements over time, but the data remain primarily self-reported and cannot establish definitive causality. Participants were recruited in China through non-probability online and organization-assisted methods, meaning the findings should not automatically be generalized to other countries, sectors or workforce cultures.

Future research needs cross-national comparisons, experimental or longer-term designs, and objective measures of how employees actually change tasks, learn new skills or withdraw from work under AI pressure. Different technologies may also generate different psychological responses: generative AI, automation systems and decision-support tools do not threaten occupations in identical ways.